A method for producing a hollow fiber, a hollow fiber, and a hollow fiber module

Abstract

The present invention relates to a method of preparing a hollow fiber, a hollow fiber, and a hollow fiber module. The present invention provides a method for preparing a hollow fiber, comprising the steps of: providing a dope solution comprising polyamide-imide (PAI); providing an aqueous bore solution comprising polyethyleneimine (PEI); extruding the dope solution in a cross-sectional ring shape and jetting the aqueous bore solution at the center of the ring shape; allowing the PAI and the PEI to react, thereby forming an inner surface layer comprising a cross-linking reaction product; and forming a polyamide thin film composite (TFC) layer by interfacial reaction of an aqueous diamine or triamine compound and an organic di- or triacyl halide compound on the inner surface layer comprising the cross-linking reaction product, such that the inner surface layer serves as a primer connecting the TFC layer to a PAI support layer. The present invention also provides a hollow fiber and a hollow fiber module.

Description

[0001] This application is a divisional application of the invention patent application filed on June 25, 2021, entitled "Hollow Fiber and Preparation Method Thereof", with application number 2021800462920. Technical Field

[0002] This disclosure relates to a hollow fiber, a method for preparing the hollow fiber, and a hollow fiber assembly comprising a bundle of hollow fibers. The hollow fiber exhibits high integration between the support membrane and the selective layer, ensuring physical stability under various conditions. Background Technology

[0003] Hollow fiber modules are gaining popularity for forward osmosis (FO). The FO process utilizes an osmotic pressure gradient instead of the hydraulic pressure used in reverse osmosis (RO) as the driving force, thus offering the potential to extract water with less energy consumption. FO is well-suited for a wide range of potential applications, including wastewater treatment, water purification, seawater desalination, food and beverage processing, and power generation.

[0004] In the food and beverage sector, WO 2019215226 has recommended the use of hollow fiber assemblies for enriching the ethanol concentration in liquids such as beer or wine. Other emerging areas include freeze-drying or spray-drying concentrated liquid coffee, concentrated coconut water, concentrated fruit juice, etc.

[0005] Hollow fiber modules are also suitable for reverse osmosis (RO) operations, such as low-pressure RO. Hollow fiber modules used in low-pressure RO operations can be used, for example, to produce drinking water from impure sources.

[0006] Hollow fibers are typically produced by hollow fiber spinning, particularly by dry-jet wet spinning as disclosed, for example, in Shi et al. (J Membrane Sci. 305 (2007) 215-225).

[0007] WO 2012074487 relates to a method for forming a hollow fiber membrane comprising a polyamide-imide (PAI) hollow fiber substrate having a PAI surface layer crosslinked with polyethyleneimine (PEI). The membrane is suitable for forward osmosis applications. The method includes contacting the PAI hollow fiber substrate, on its outer side, with an aqueous solution of polyethyleneimine (PEI) under conditions suitable for crosslinking the PAI in the surface layer of the PAI hollow fiber substrate with PEI.

[0008] The objective is to provide a robust hollow fiber with high physical integrity. In particular, the objective is to achieve high adhesion between the support membrane and the selective layer, thereby avoiding or reducing delamination. Furthermore, the objective of this invention is to provide a membrane with fewer support defects, i.e., defects in the selective layer caused by the surface topology of the hollow fiber. Summary of the Invention

[0009] In a first aspect, the present invention relates to a method for preparing hollow fibers, comprising the following steps: a) Provide coating solutions including polyamide-imide (PAI), b) Provide an aqueous porous solution including polyethyleneimine (PEI), c) Extruding the coating solution in a cross-sectional annular shape and spraying the orifice solution at the center of the annular shape. d) Allowing the PAI and PEI to react, thereby forming an inner surface layer comprising the crosslinking reaction product, and e) A polyamide film composite (TFC) layer is formed by interfacial reaction of an aqueous diamine or triamine compound and an organic diacyl halide or triacyl halide compound on the inner surface layer including the crosslinking reaction product.

[0010] Adding PEI to the aqueous porous solution allows for the formation of an inner surface layer, also known as a gutter layer, within the lumen (inner surface) of the hollow fiber membrane by reacting the PAI with the PEI. This inner surface layer serves as a primer for attaching the TFC layer to the PAI support layer. Therefore, the presence of this inner surface layer ensures high adhesion and prevents delamination of the TFC layer. Furthermore, PEI contains amine groups, and it is assumed that these groups, along with diamine and triamine compounds, participate in the formation of the TFC layer, thereby forming a covalent bond between the TFC layer and the PAI support layer through the gutter layer.

[0011] In some implementations of the invention, the transition layer tends to make the surface roughness, often observed on the inner surface of the hollow fiber, uniform. Therefore, the smooth surface provided by the transition layer reduces support defects, i.e., defects in the TFC layer caused by height variations in the surface topology.

[0012] In some embodiments of the invention, a smooth transition layer provides a slower diffusion rate of the diamine or triamine compound during the formation of the interfacial reaction, ensuring a thinner and / or denser TFC layer.

[0013] According to a second aspect, a hollow fiber is provided, comprising an outer support layer, an inner surface layer, and a polyamide film composite (TFC) layer, wherein the outer support layer comprises polyamide-imide (PAI), the inner surface layer comprises a reaction product of PAI and polyethyleneimine (PEI), and the TFC layer comprises a reaction product of a diamine or triamine compound and a diacyl halide or triacyl halide compound.

[0014] According to a third aspect of the invention, a hollow fiber assembly is provided, comprising a bundle of hollow fibers as described in the second aspect.

[0015] In a possible implementation of the first aspect, the monomer of the PAI has the following chemical structure:

[0016] Ar is selected from the following aromatic groups: , , or .

[0017] In a possible implementation of the first aspect, the coating solution further comprises an aprotic solvent.

[0018] In a possible implementation of the first aspect, the aprotic solvent is selected from the group consisting of N-methylpyrrolidone (NMP), dimethylacetamide (DMAC), dimethyl sulfoxide (DMSO), and dimethylformamide (DMF), or mixtures thereof.

[0019] In a possible implementation of the first aspect, the coating solution further comprises a polymer selected from the group consisting of: polyphenylene sulfone (PPSU), polyether sulfone (PES), sulfonated polyether sulfone (SPES), polysulfone (PSU), polyether ether ketone (PEEK), high-temperature sulfone resin, self-reinforced polyphenylene, polybenzimidazole (PBI), polyimide (PI), polyether imide (PEI), polyphenylene sulfide (PPS), and mixtures thereof.

[0020] In a possible implementation of the first aspect, the PAI has a solution viscosity of 30,000 cP or less when measured in 25% NMP at 40°C.

[0021] In one possible implementation of the first aspect, the coating solution further comprises a diol.

[0022] In a possible implementation of the first aspect, the diol is selected from ethylene glycol, 1,2-propanediol, 1,3-propanediol, glycerol, or polyethylene glycol (PEG). Typically, the PEG has a molecular weight of 10,000 D or lower, such as PEG200 or PEG400.

[0023] In a possible implementation of the first aspect, the aqueous pore solution has a PEI of 2,000 g / mol or more, such as 25,000 g / mol or more, and preferably an average molecular weight (M) of 100,000 g / mol or more. W ).

[0024] In one possible implementation of the first aspect, the PEI is branched.

[0025] In a possible implementation of the first aspect, the aqueous porous solution further comprises an aprotic solvent.

[0026] In a possible implementation of the first aspect, the aprotic solvent is selected from the group consisting of N-methylpyrrolidone (NMP), dimethylacetamide (DMAC), dimethyl sulfoxide (DMSO), and dimethylformamide (DMF), or mixtures thereof.

[0027] In one possible implementation of the first aspect, step c) is carried out by dry-jet wet spinning.

[0028] In one possible implementation of the first aspect, the initial hollow fibers after extrusion move through the air gap before solidifying in a non-solvent.

[0029] In one possible implementation of the first aspect, the air gap is 5 to 40 cm.

[0030] In one possible implementation of the first aspect, a plurality of hollow fibers are assembled in the assembly prior to the preparation of the polyamide TFC layer.

[0031] In a possible implementation of the first aspect, the diamine or triamine compound is selected from m-phenylenediamine (MPD), diethylenetriamine, dipropylenetriamine, phenyltriamine, bis(hexamethylene)triamine, bis(hexamethylene)triamine, bis(3-aminopropyl)amine, hexamethylenediamine, N-tartrate dipropylene, 1,3,5-triazine-2,4,6-triamine, and mixtures of these compounds.

[0032] In a possible implementation of the first aspect, the diacyl halide or triacyl halide compound is selected from trimesoyl chloride (TMC), trimesoyl bromide, isophthaloyl chloride (IPC), isophthaloyl bromide, terephthaloyl chloride (TPC), terephthaloyl bromide, adipicoyl chloride, cyanuric chloride, and mixtures thereof.

[0033] In one possible implementation of the first aspect, aquaporin water channels are introduced into the TFC layer.

[0034] In one possible implementation of the first aspect, the aquaporin water channels are assembled in a nanostructure comprising polyethyleneimine.

[0035] In a possible implementation of the first aspect, the polyethyleneimine has an average molecular weight of about 2,000 Da to about 10,000 Da, such as about 3,000 Da to about 5,000 Da.

[0036] In one possible implementation of the first aspect, the aquaporin water channels are dissolved in a cleaning agent before being assembled in a nanostructure comprising polyethyleneimine.

[0037] In a possible implementation of the first aspect, the cleaning agent is selected from the group consisting of lauryl dimethylamine N-oxide (LDAO), octyl glucoside (OG), dodecyl maltodextrin (DDM), or combinations thereof.

[0038] In one possible implementation of the first aspect, the aquaporin water channels are provided in vesicles before being introduced into the TFC layer.

[0039] In a possible implementation of the first aspect, the vesicle comprises an amphiphilic diblock copolymer of the PMOXA-PDMS type and reactive end-functionalized PDMS.

[0040] In a possible implementation of the first aspect, the PMOXA-PDMS is selected from PMOXA. 10-40 -PDMS 25-70 A group consisting of its mixtures.

[0041] In a possible implementation of the first aspect, the mixture comprises at least the general formula PMOXA 10-28 -PDMS 25-70 The first amphiphilic diblock copolymer and general formula PMOXA 28-40 -PDMS 25-70 The second amphiphilic diblock copolymer.

[0042] In a possible implementation of the first aspect, the reactive end-functionalized PDMS is functionalized by one or more of an amine, carboxylic acid, and / or hydroxyl groups.

[0043] Among the possible implementations of the first aspect, a triblock copolymer of the PMOXA-PDMS-PMOXA type is further included.

[0044] In one possible implementation of the first aspect, the TFC layer exists inside the fiber.

[0045] In a possible implementation of the second aspect, the outer support layer further comprises a polymer selected from the group consisting of: polyphenylene sulfone (PPSU), polyether sulfone (PES), polysulfone (PSU), polyether ether ketone (PEEK), high-temperature sulfone resin, self-reinforced polyphenylene, polybenzimidazole (PBI), polyimide (PI), polyether imide (PEI), polyphenylene sulfide (PPS), and mixtures thereof.

[0046] In a possible implementation of the second aspect, the PAI has a solution viscosity of 30,000 cP or less when measured in 25% NMP at 40°C.

[0047] In a possible implementation of the second aspect, the PEI has an average molecular weight (M W The concentration is 100,000 g / mol or higher.

[0048] In a possible implementation of the second aspect, the PEI is branched.

[0049] In a possible implementation of the second aspect, the diamine or triamine compound is selected from m-phenylenediamine (MPD), diethylenetriamine, dipropylenetriamine, phenyltriamine, bis(hexamethylene)triamine, bis(hexamethylene)triamine, bis(3-aminopropyl)amine, hexamethylenediamine, N-tartrate dipropylene, 1,3,5-triazine-2,4,6-triamine, and mixtures of these compounds.

[0050] In a possible implementation of the second aspect, the diacyl halide or triacyl halide compound is selected from trimesoyl chloride (TMC), trimesoyl bromide, isophthaloyl chloride (IPC), isophthaloyl bromide, terephthaloyl chloride (TPC), terephthaloyl bromide, adipicoyl chloride, cyanuric chloride, and mixtures thereof.

[0051] In a possible implementation of the second aspect, the aquaporin water channels are introduced into the TFC layer.

[0052] In a possible implementation of the second aspect, the aquaporin water channels are assembled in a nanostructure comprising polyethyleneimine.

[0053] In a possible implementation of the second aspect, the polyethyleneimine has an average molecular weight of about 2,000 Da to about 10,000 Da, such as about 3,000 Da to about 5,000 Da.

[0054] In a possible implementation of the second aspect, the aquaporin water channels are dissolved in a cleaning agent before being assembled in a nanostructure comprising polyethyleneimine.

[0055] In a possible implementation of the second aspect, the cleaning agent is selected from the group consisting of lauryl dimethylamine N-oxide (LDAO), octyl glucoside (OG), dodecyl maltodextrin (DDM), or combinations thereof.

[0056] In a possible implementation of the second aspect, the aquaporin water channels are provided in vesicles.

[0057] In a possible implementation of the second aspect, the vesicle comprises an amphiphilic diblock copolymer of the PMOXA-PDMS type and reactive end-functionalized PDMS.

[0058] In a possible implementation of the second aspect, the PMOXA-PDMS is selected from PMOXA. 10-40 -PDMS 25-70 A group consisting of its mixtures.

[0059] In a possible implementation of the second aspect, the mixture comprises at least the general formula PMOXA 10-28 -PDMS 25-70 The first amphiphilic diblock copolymer and general formula PMOXA 28-40 -PDMS 25-70 The second amphiphilic diblock copolymer.

[0060] In a possible implementation of the second aspect, the reactive end-functionalized PDMS is functionalized by one or more of an amine, carboxylic acid, and / or hydroxyl groups.

[0061] In a possible implementation of the second aspect, a triblock copolymer of the PMOXA-PDMS-PMOXA type is further included.

[0062] In a possible implementation of the second aspect, the hollow fiber has a wall thickness of 0.10 ± 0.01 mm or more.

[0063] In a possible implementation of the second aspect, the inner diameter of the hollow fiber is 0.80 ± 0.01 mm or more.

[0064] In a possible implementation of the second aspect, the hollow fiber has a porosity of 75% or more.

[0065] In a possible implementation of the third aspect, the hollow fiber assembly includes a housing surrounding a bundle of hollow fibers, an inlet and an outlet for allowing a first solution to pass through the inner cavity side of the assembly, connected to the inner cavity of the hollow fibers at each end of the bundle, wherein the inlet and outlet are provided in the housing for allowing a second solution to pass through the housing side of the assembly.

[0066] In a possible implementation of the third aspect, the bundle of hollow fibers is encapsulated in each end of the component to form a liquid-tight seal between the outer surface of the hollow fibers and the housing.

[0067] These and other aspects will be apparent from one or more embodiments described below. Attached Figure Description

[0068] Figure 1 SEM images of bare fibers and M0.12 fibers are shown: a) cross-section, b) inner surface before IP coating, and c) inner surface after IP coating.

[0069] Figure 2 The burst pressure of the fiber support is shown in contrast to the wall thickness (t) normalized to OD (outer diameter), n=5.

[0070] Figure 3 The zeta potential of the hollow fiber support varies with pH (bars represent the standard deviation of 3 fibers, with 6 measurements for each).

[0071] Figure 4 The permeability of the fiber support to pure water (before IP coating) is shown, n=3.

[0072] Figure 5A The figure shows that the NaCl rejection rate varies with water permeability (A), and Figure 5B The figure shows that the TSC retention rate varies with (A), n=3.

[0073] Figure 6A The figure shows the water flux (Jw) and reverse salt flux (Js) measured in FO mode using DI water as the feed solution and 1M NaCl as the draw solution. Figure 6B The figure shows the water flux (Jw) and reverse salt flux (Js) measured in FO mode using DI water as the feed solution and 1M TSC as the draw solution, n=3.

[0074] Figure 7A and Figure 7B The S-parameters are shown, measured in FO mode and calculated using Equation 5. Figure 7A 1M NaCl and Figure 7B :1M TSC, n=3.

[0075] Figure 8 SEM images of unmodified fiber compared to 750K fiber are shown: (a) cross-section, (b) morphology of the middle portion, (c) inner surface before IP, and (d) inner surface after IP.

[0076] Figure 9 A cross-sectional image of the newly spun fiber (750K) is disclosed, compared with the fiber (M0.12) spun with 3 wt% polymer in Section 5.

[0077] Figure 10 The zeta potential of the inner surface of the HF support as a function of pH was depicted, n=12. (M0.12 was spun with an additional 3 wt% polymer content, and the data were taken from Section 5 for comparison).

[0078] Figure 11 The PWP of the support, measured from the inside out in a dead-end configuration at 1 bar, is shown, n=3. (M0.12 was spun with an additional 3 wt% polymer content, and the data are taken from Section 5 for comparison).

[0079] Figure 12 Separate data are shown for the NaCl rejection of fibers after IP coating with 500 mg·L⁻¹ NaCl solution at 2 bar, compared with their water permeability measured with DI water at 2 bar. (M0.12 was spun with a polymer content of more than 3 wt%, and the data are taken from Section 5 for comparison).

[0080] Figure 13 The permeation performance of an HF membrane using 1M NaCl as the draw solution and DI water as the feed solution in FO mode is disclosed. Left y-axis / bar: water flux. Right y-axis / square: reverse salt flux, n=3. (M0.12 was spun with a polymer content of more than 3 wt%, and the data are taken from Section 5 for comparison).

[0081] Figure 14 The permeation performance of the HF membrane in PRO mode using 1M NaCl as the draw solution and DI water as the feed solution is shown. Left y-axis / bar: water flux. Right y-axis / square: reverse salt flux, n=3. (M0.12 was spun with an additional 3 wt% polymer content, and the data are taken from Section 5 for comparison).

[0082] Figure 15 The structural parameters of the HF membrane, n=3, are shown in FO and PRO modes using 1M NaCl as the draw solution and DI water as the feed solution. Detailed Implementation

[0083] Hollow fibers are typically produced using a process called dry-jet wet spinning, in which a coating solution and a porous fluid are extruded together through a spinneret. The coating solution is extruded through a ring-shaped, dual-orifice spinneret, and the porous fluid is extruded through a central, ring-shaped spinneret. The dual-orifice spinneret includes an outer annular die for the coating solution and a central die for the porous fluid. The spun hollow fibers enter a coagulation bath after passing through a defined air gap distance. In the coagulation bath, solid hollow fibers are formed through solvent-induced phase separation.

[0084] The PAI in the initial hollow fibers is modified after extrusion by reacting with PEI present in the pore solution. More specifically, the carbonyl groups of PAI react with the amine groups of PEI to form a cross-linked polymer. After a certain reaction time, the initial hollow fibers are solidified in a coagulation bath.

[0085] The PAI support layer, modified with the internal crosslinking reaction product between PAI and PEI, is further modified by interfacial polymerization to form a thin film composite (TFC) layer. The TFC layer can be prepared using amine reactants in an aqueous solution, preferably aromatic amines, such as diamines or triamines, for example 1,3-diaminobenzene (m-phenylenediamine-MPD), and acyl halide reactants dissolved in an organic solvent, such as diacyl chlorides or triacyl chlorides, preferably aromatic acyl halides, such as benzene-1,3,5-tricarbonyl chloride (TMC), wherein the reactants are combined in an interfacial polymerization reaction.

[0086] The amine groups of the diamine or triamine compounds and the amine groups of PEI will compete with the acyl halide groups of the acyl halide compounds for reaction. Therefore, it is believed that the TFC layer is covalently bonded to the PAI support film through the transition layer.

[0087] PAI is a thermoplastic amorphous polymer with properties derived from the positive synergistic effect of both polyamides and polyimides, resulting in excellent mechanical, heat, and chemical resistance. PAI exhibits good chemical stability over a wide pH range and demonstrates high resistance to many organic solvents due to its ability to form intra- and inter-chain hydrogen bonds. Furthermore, PAI can be used to manufacture hollow fiber membranes using non-solvent-induced phase separation (NIPS) technology. PAI is primarily manufactured by Solvay Advanced Polymers under the trademark Torlon (Torlon®). Three types of Torlon® PAI materials are particularly useful in membrane manufacturing: Torlon® 4000T-LV (low viscosity), Torlon® 4000T-MV (medium viscosity), Torlon® 4000T-HV (high viscosity), and a finer-grained version, Torlon® 4000TF. In one embodiment, Torlon® 4000T-TF is used in this invention. Torlon® PAI powder is soluble in dipolar aprotic solvents such as N-methylpyrrolidone (NMP), dimethylacetamide (DMAC), dimethyl sulfoxide (DMSO), and dimethylformamide (DMF). Torlon® 4000T-TF has the appearance of a fine yellow powder and a solution viscosity of 7,000 cP in 25% NMP measured at 40°C.

[0088] While the membrane is expected to be used in any of the aforementioned semipermeable membranes and to facilitate forward osmosis, water flux typically becomes more efficient when aquaporin water channels are introduced into the TFC layer. Aquaporin water channels are widely distributed transmembrane proteins in nature, used to selectively transport water into or out of cells. In industrial environments, aquaporin water channels in semipermeable membranes ensure water flow through osmosis, while other solutes in the solution are retained. Therefore, the presence of active aquaporin water channels helps semipermeable membranes retain solutes and facilitates water permeation through the membrane.

[0089] Aquaporin water channels are introduced into membranes with an active structure for at least a significant amount of molecules. According to aspects of the invention, the activity of aquaporin water channels is maintained when they are assembled in nanostructures comprising polyalkylene imides (such as polyethyleneimine). As further explained in detail in WO17137361 (which is incorporated herein by reference in its entirety), polyalkylene imides, such as polyethyleneimine (PEI), form self-assembled nanostructures with transmembrane proteins, such as aquaporin water channels. The nanostructures ensure that at least a portion of the aquaporin water channels remain active even after introduction into a TFC layer. It is believed that the polymer interacts with the transmembrane protein to prevent it from reacting with monomers involved in the formation of the TFC layer. Furthermore, it is believed that the PEI of the aquaporin nanoparticles reacts with the PAI and thus becomes integrated into the transition layer.

[0090] Generally, PEI is essentially a linear or branched polymer with an average molecular weight of about 2,000 Da to about 10,000 Da, such as about 3,000 Da to about 5,000 Da. It is currently believed that the relatively short polymer interacts with transmembrane proteins to prevent them from reacting with monomers involved in TFC layer formation, while at the same time it does not substantially inhibit interactions with water.

[0091] To prevent the aggregation of aquaporin water channels, it can be advantageous to dissolve the aquaporin water channels in a cleaning agent before assembly into nanostructures comprising polyalkylimides. Since aquaporin water channels are naturally present in cell membranes, the protein exhibits hydrophobic domains. It is believed that the hydrophobic domains of the cleaning agent interact with the hydrophobic domains of the aquaporin water channels, thereby forming dissolved proteins. Although aquaporin water channels can be dissolved by a variety of cleaning agents, those selected from the group consisting of LDAO, OG, DDM, or combinations thereof are currently preferred.

[0092] In another embodiment of the invention, the aquaporin water channels are provided in vesicles prior to the introduction of the TFC layer. Vesicles are the natural environment of the aquaporin water channels, and vesicles can be formed from many different membrane-forming materials, including naturally occurring phospholipids. In one embodiment of the invention, vesicles are formed from amphiphilic diblock copolymers, such as poly(2-methyloxazoline)-block-poly(dimethylsiloxane) diblock copolymers (PMOXA-PDMS) and reactive end-functionalized poly(dimethylsiloxane) (PDMS). As further explained in detail in WO 2018 / 141985 (which is incorporated herein by reference in its entirety), the above components self-assemble into vesicles within the liquid composition.

[0093] The two blocks of a PMOXA-PDMS diblock copolymer can have different lengths. To obtain sufficient vesicle stability, the PMOXA-PDMS diblock copolymer is typically selected from PMOXA... 10-40 -PDMS 25-70 The group consists of mixtures thereof. Experiments have shown that mixtures of different PMOXA-PDMS diblock copolymers exhibit higher robustness. In a preferred embodiment, the vesicles therefore comprise at least the general formula PMOXA. 10-28 -PDMS 25-70 The first amphiphilic diblock copolymer and general formula PMOXA 28-40 -PDMS 25-70 The second amphiphilic diblock copolymer. The weight ratio between the first and second amphiphilic diblock copolymers is typically in the range of 0.1:1 to 1:0.1. The concentration of the amphiphilic diblock copolymer in the liquid composition is typically in the range of 0.1 to 50 mg / ml, such as 0.5 to 20 mg / ml, and preferably 1 to 10 mg / ml.

[0094] The reactively terminal-functionalized PDMS (reactively terminal-functionalized poly(dimethylsiloxane)) of the vesicles can be functionalized with one or more of amine, carboxylic acid, and / or hydroxyl groups. In one aspect of the invention, the reactively terminal-functionalized PDMS... e-f It is a PDMS with bis(aminoalkyl), bis(hydroxyalkyl), or bis(carboxyalkyl) end caps. e-f Examples include poly(dimethylsiloxane), bis(3-aminopropyl), or poly(dimethylsiloxane), bis(3-hydroxypropyl). Suitably, the integer e is selected from the range of 20 to 40, such as 30, and the integer f is selected from the range of 40 to 80, such as 50. Furthermore, reactive end-functionalized PDMS... e-f Optional H2N-PDMS 30-50 HOOC-PDMS 30-50 and HO-PDMS 30-50 The group consists of mixtures thereof. Before the vesicles are introduced into the aquaporin water channels, the vesicles may be present in the liquid composition, and the amount of PDMS is preferably from about 0.05% to about 1% v / v.

[0095] The vesicles of the present invention may further comprise about 1% v / v to about 12% v / v of PMOXA. a-b -PDMS c-d -PMOXA a-b The vesicles are composed of triblock copolymers to increase their integrity. Typically, the vesicles comprise approximately 8% v / v to approximately 12% v / v of PMOXA. a-b -PDMS c-d -PMOXA a-bPMOXA is a type of triblock copolymer. a-b -PDMS c-d -PMOXA a-b Triblock copolymers of this type are typically selected from PMOXA. 10-20 -PDMS 25-70 -PMOXA 10-20 .

[0096] The vesicles of the present invention may further include flux enhancers to increase water flux or reduce reverse salt flux. Flux enhancers may be selected from a large group of compounds, typically preferably alkylene glycol monoalkyl ether alkylates, β-cyclodextrins, or polyethylene glycol (15)-hydroxystearate. The flux enhancer is typically present in an amount of 0.1% to 1% by weight of the liquid composition.

[0097] The vesicles of the present invention may be present in a liquid composition prior to being immobilized in a membrane (such as a TFC layer provided on a support membrane). The liquid composition may include a buffer to stabilize the vesicles. The transmembrane protein is appropriately dissolved in a cleaning agent before being mixed with other components. The vesicles in the liquid composition may further include a cleaning agent or a surfactant. The cleaning agent may be selected from the group consisting of lauryl dimethylamine N-oxide (LDAO), octyl glucoside (OG), dodecyl maltodextrin (DDM), or combinations thereof.

[0098] Without being bound by any particular theory, it is believed that vesicles containing free, available reactive groups on the surface will not only be physically introduced or immobilized (adsorbed) but also chemically bonded to the TFC layer, as reactive free end groups such as amino, hydroxy, and carboxyl groups will participate in interfacial polymerization reactions with acyl chlorides such as trimesoyl chloride (TMC). In this way, vesicles are believed to be covalently bound to the TFC layer, resulting in a relatively high vesicle loading and therefore a higher water flux through the membrane. Furthermore, it is currently believed that free end groups, such as amino or hydroxyl groups, can react with the carbonyl groups of PAI to form covalent bonds between the vesicles and the supporting hollow fiber membrane. Moreover, it is believed that when introduced into selective membrane layers, the covalent coupling of vesicles in the TFC layer leads to greater stability and / or lifetime of aquaporin water channels and vesicles containing aquaporin water channels.

[0099] Vesicles can be prepared in liquid compositions incorporating aquaporin water channels, including stirred PMOXA. a-b -PDMS c-d A solution of a type of amphiphilic diblock copolymer, and 0.05% to about 1% of reactive end-functionalized PDMS. e-f The steps involve mixing the aquaporin aquaporin with the aquaporin aquaporin. For best results, stirring should continue for 12-16 hours.

[0100] The preparation of a thin-film composite layer immobilizing vesicles introducing aquaporin water channels on a porous basement membrane comprises the following steps: providing a mixture of vesicles and a diamine or triamine compound in a liquid composition as disclosed above; covering the surface of a porous support membrane with the mixture; applying a hydrophobic solution comprising an acyl halide compound; and allowing the aqueous solution and the hydrophobic solution to undergo an interfacial polymerization reaction to form the thin-film composite layer. In one embodiment of the invention, the hydrophobic solution further comprises 0.1 to 10% by volume of a TFC layer modifier. The purpose of the TFC layer modifier is to increase water flow and / or solute retention. In suitable embodiments, the TFC layer modifier is a C3 to C8 carbonyl compound. As an example, the TFC layer modifier is selected from the group consisting of: diethylidene ketone, 2-pentanone, 5-pentanone, and / or cyclopentanone.

[0101] Example Example 1 1.1.1 Materials Polyamide-imide (PAI) Torlon® 4000TF was supplied amicably by Solvay Advanced Polymers and used as the base polymer for coating solution preparation. 1-Methyl-2-pyrrolidone (NMP) 99% was supplied by Acros Organics. Isopar E was received amicably from ExxonMobil Petroleum & Chemical BVBA (Belgium). 1,3-Phenylenediamine (MPD) >98%, ethylene glycol (EG), glycerol 85-87%, 1,3,5-pyromellitic acid chloride (TMC) >98%, trisodium citrate (TSC), and polyethyleneimine (PEI, Mw ~ 750,000 g·mol⁻¹) -1 The 50 wt% aqueous solution was purchased from Sigma-Aldrich. All chemicals were used as is without further purification.

[0102] 1.1.2 Manufacturing of hollow fibers Following several development steps, a coating composition consisting of 18 wt% PAI, 13 wt% EG, and 69 wt% NMP was selected. During hollow fiber spinning, we modified the inner lumen side of our hollow fibers with PEI in a one-step process to enhance subsequent interfacial polymerization steps.

[0103] Before preparing the coating solution, the base polymer (PAI) was dried in a drying oven at 100°C for 24 h. After adding all compounds to a mixing flask, both the pore solution and the coating solution were prepared using a roller table. The coating was degassed by placing the polymer solution in the coating container of a spinning machine for 24 h. To investigate the effect of PEI modification, two different aqueous pore solutions with and without PEI were used (Table 1).

[0104] Table 1. Pore solution compositions.

[0105]

[0106] In total, seven different fibers were spun using dry-jet wet spinning technology (Table 2). The first batch (bare) was manufactured using pore solution 1 (B1) as a reference for further analysis, while the other six fibers were manufactured using pore solution 2 (B2) and by varying spinning conditions to achieve different wall thicknesses. The freshly spun fibers were rinsed with deionized (DI) water for 48 hours and then immersed in a 30% w / w glycerol aqueous solution for 24 hours before drying.

[0107] Table 2. Spinning conditions.

[0108]

[0109] 1.1.3 Component Preparation For the filtration experiments, five fibers were placed into a 10 mm perfluoroalkoxy (PFA) tube each time, with two QST-10 Festo push-in fittings attached to each side. Both ends were sealed with slow-curing epoxy resin for a length of 4.5 cm to achieve a fiber length of 16 cm. Three sets were prepared for further experiments for each type of fiber.

[0110] 1.2 Membrane Characterization 1.2.1 Scanning Electron Microscopy Hollow fibers were first frozen in liquid nitrogen and then fractured to prepare cross-sectional samples. For analysis of the inner cavity surface, the fibers were diagonally cut with a sharp blade. All samples were stored in a vacuum oven at 30°C for 24 h. Finally, a 10 nm chromium layer was sputtered onto the sample surface using a Quorum Q150T ES sputtering coating machine. Cross-sectional and inner cavity surface images of the samples were obtained using a JEOL JSM-6010LA scanning electron microscope.

[0111] 1.2.2 Mechanical Strength The burst pressure of the single fiber was determined using a custom setup manufactured by Convergence Industry BV, Netherlands. It forces fluid through the hollow fiber under increasing pressure until the sample fails. This ultimate pressure was reported. For this purpose, a single-fiber membrane module was fabricated using a material similar to that used in the filtration experiments. Pressure was applied from the inside of the membrane in increments of 0.5 bar per second. When using the fluid glycerol, less permeation is allowed due to its high viscosity, and therefore pressure control is easier. The experiments were conducted at room temperature.

[0112] Under ultimate pressure, the tensile strength (δ) of the fiber is determined by Barlow's Formula, given in Equation 2.

[0113] (2) Where P is the obtained burst pressure (bar), t is the wall thickness (mm), and D... 外 δ is the outer diameter (mm) and δ is the tensile strength (bar).

[0114] 1.2.3 Porosity Measurement Porosity was measured using the known value of the polyamide-imide density. Before measuring its weight, the sample was placed in a vacuum oven for at least 1 hour. Porosity was calculated using Equation 3.

[0115] (3) m is the sample mass (g), and ρ is the density of PAI (g·cm³). -3 V is the volume of the fiber (cm³). -3 The density of the polyamide-imide used in this calculation is 1.4 g·cm³. -3 Each type of fiber was tested in triplicate.

[0116] 1.2.4 Determination of ζ potential The zeta potential of the fibers was determined at different pH values ​​(3-11) using a 5 mM KCl electrolyte solution and an electric SurPASS analyzer (Anton Paar, Graz, Austria). In addition to bare fibers and one-step modified fibers, post-modified fibers (Post-M) with PEI were tested. To prepare Post-M fibers, three bare fibers were immersed in a 1 wt% PEI solution at pH 5.8 for 12 h. This was done to demonstrate the surface charge of pure PEI for comparative purposes.

[0117] For zeta potential measurements, three single fibers of each type were encapsulated in 7 cm long tubes for triplicate measurements. The pH was automatically adjusted using two solutions: 0.1 M HCl and 0.1 M NaOH. The zeta potential was calculated using Equation 4.

[0118] (4) Here, ζ is the zeta potential (V), I is the flowing current (A), P is the pressure (Pa), η is the dynamic viscosity of the electrolyte solution (Pa·s), ε is the dielectric permittivity of water, and ε0 is the dielectric permittivity in vacuum (F·m). -1 ), L s It is the channel length (m), and A s It is the cross-section of the flow channel (m) 2 ).

[0119] 1.2.5 Measurement of Pure Water Permeability The pure water permeability (PWP) of the fiber was determined using nitrogen gas at a pressure of 2 bar in a dead-end, inside-out configuration. -2 h -1 -bar -1 ).

[0120] (5) Here, V is the volume of permeate collected (L), t is the collection time (h), and A is the membrane inner surface area (m²). 2 ),as well as P is the pressure difference (bar) between the inner cavity and the outer shell of the hollow fiber.

[0121] 1.2.6 Interface Aggregation Interfacial polymerization (IP) was performed on the inner surface of the fibers using MPD and TMC to form a polyamide (PA) selective layer. To achieve adequate wetting and removal of glycerol from the fibers, the components were immersed overnight in DI water prior to the IP reaction. IP coating was performed using a 2.5% (w / v) aqueous solution of MPD and a 0.15% (w / v) solution of TMC. TMC was dissolved in isopar E using an ultrasonic bath (20°C, 2 min).

[0122] First, excess water in the shell and hollow fiber cavity is removed by passing N2 stream for 1.5 min. Then, water is injected at 10 L / min using a syringe pump. -1 The MPD solution was fed into the inner cavity of vertically fixed fibers (5 fibers) for 2 min. Then, the MPD solution was kept horizontal within the fibers for 10 min, followed by removal of excess MPD by passing it through a stream of N2 for 1 min. Subsequently, the solution was injected using a syringe pump at a rate of 10 L / min.-1 The TMC solution was introduced into the vertically fixed fibers twice for 15 seconds, with a 45-second interval. Residual TMC was then removed by passing it through a stream of N2 for 30 seconds, followed by drying in an air oven at 70°C for 10 minutes. Finally, the fiber cavities were rinsed with 20 ml of DI water, and the assembly was kept in DI water at 7°C until further use.

[0123] 1.2.7 Salt Retention Rate Use 500 mg·L -1 The salt rejection of coated fibers was determined using NaCl and TSC solutions in an inside-out cross-flow filtration system at a transmembrane pressure of 2 bar. The cross-flow velocity per fiber was set to 1.7 m / s. -1 The rejection rate is determined by measuring the conductivity of the feed solution and the collected permeate, and then calculated as follows: (6) In this equation, R is the salt rejection percentage (%), and σ f The conductivity of the feed solution (μS·cm) -1 ), and σ p It is the conductivity of the permeate solution (μS·cm) -1 ).

[0124] Based on the Lévêque interrelationships used for the fully developed laminar velocity profile within the pipe, J w (L·m -2 h -1 Water flux and mass transfer coefficient (K) (L·m) -2 h -1 Determine salt permeability (B).

[0125] (7) 1.2.8 FO Performance The permeation performance of fibers was evaluated in a selective layer in a feed-oriented (FO) mode. Tests were conducted in a countercurrent configuration using either 1 M NaCl or 1 M TSC as the draw solution, with a cross-flow rate of 3 L / h for each component. -1 (corresponding to 2 m·s in the inner cavity) -1 The experiment lasted for 1 hour, during which the water flux (J) was determined by measuring the amount of water transferred from the feed solution to the draw solution per membrane area each time. w (L·m) -2 h -1 Reverse salt flux (J) s (g·m) -2 h -1 The method is determined by measuring the increase in conductivity of the feed solution per membrane area each time. The amount of solute passing through the feed side is measured using a pre-defined calibration curve.

[0126] Since the curvature is difficult to measure, the structural parameter (S) was estimated by fitting, which can be written as (Equation 8) for the FO model: (8) Here, J w (L·m -2 h -1 D(m) is the infiltration flow rate. 2 ·h -1 π is the solute diffusivity. D (bar) is the osmotic pressure of the solution drawn, and A is the water permeability (L·m). -2 h -1 -bar -1 B is the salt permeability coefficient (L·m) determined by Equation 7. -2 h -1 ).

[0127] For all the results presented, the presented error is the standard deviation of the obtained measurements.

[0128] 1.3 Results and Discussion 1.3.1 Hollow fiber size and porosity The thickness, inner diameter, and porosity of the spun fibers can be found in Table 3. The fiber wall thickness varied from 0.07 to 0.17 mm. The measured porosity of the single-step modified fibers ranged from 75% to 77%. Based on Equation 1, a 2% change in porosity leads to a 2% change in S-parameters, while the wall thickness difference (0.07–0.17 mm) used in this study affects S-parameters by 143%. This leads us to conclude that the effect of porosity in this study is very small and negligible compared to wall thickness. Furthermore, we analyzed the effect of the porosity of these fibers on S-parameters (determined by Equation 7), and no real interrelationship was found in (data not reported here).

[0129] Table 3. Dimensions and porosity of spun hollow fiber supports determined by SEM dimensional tools, n=3.

[0130]

[0131] 1.3.2 Hollow Fiber Morphology For illustrative purposes, SEM images of bare fibers and M0.12 fibers are shown in this section because they have similar wall thicknesses. Cross-sectional and inner surface images before and after IP coating are available in [images missing]. Figure 1 Cross-sectional images of the bare fibers and M0.12 fibers show similar morphologies. They consist of finger-like structures extending from the center to the inner lumen side, with several large voids adjacent to the shell side. This shows that adding PEI to the pore solution has no effect on the overall structure of the HF support, so similar fiber curvature is expected. This is advantageous because the overall structure of the support can be further improved independently to meet certain requirements while keeping the inner surface unchanged. Nevertheless, the inner surface of the modified fiber (image M0.12-b) exhibits a slightly different structure from the bare fiber (image bare-b). In fact, it appears smoother than the inner lumen surface of the bare fiber. Similar structures can be observed for the inner surfaces of other modified fibers with different thicknesses. This similarity in the inner surface images of all modified fibers shows that the modification of the inner surface by PEI is reproducible regardless of variations in support thickness. This is important because, in order to focus only on support thickness, the pore surfaces should be similar, as IP is known to be sensitive to the physicochemical properties of the support surface, such as porosity, pore size, and hydrophilicity.

[0132] SEM images of the inner surface after IP clearly show the presence of a polyamide layer, but with a disputedly different morphology (bare -c and M0.12-c images), which can be attributed to surface modification via PEI. The addition of PEI results in a smoother inner cavity surface, which leads to a thinner PA layer. This causes the reaction to occur primarily within the pores rather than on the surface, unlike the bare fibers. For the bare fibers, the MPD solution diffuses more easily out of the surface pores during the IP process, resulting in the interface between MPD and TMC being primarily located on the surface. This results in a more pronounced ridge and valley PA layer in the bare fibers (bare -c). The same PA structure was observed for other PEI-modified fibers with different thicknesses (data not reported here), again demonstrating that the inner surface of the modified fibers remains consistent even when the wall thickness is altered.

[0133] 1.3.3 Mechanical Strength The burst pressure of the modified fibers was plotted against their outer diameter (OD) normalized thickness. Figure 2As predicted by Equation 2 (Barlow's formula), a significant upward trend in fiber burst pressure was observed with increasing wall thickness. This finding also indicates that reducing wall thickness to increase permeate flux is accompanied by sacrificial support mechanical strength. Nevertheless, a high burst pressure of 13 ± 1.5 bar was found for the thinnest fiber with a wall thickness of 0.07 mm, attributed to the high tensile strength of PAI. We believe that a burst pressure of 10 bar would be sufficient for membranes used in FO, similar to the burst pressure of commercial ultrafiltration membranes, where operating pressures are typically below 1 bar. Therefore, all membranes are mechanically suitable for the FO process.

[0134] Furthermore, the burst pressure of PAI fibers was compared with that of commercial PES / PVP ultrafiltration and microfiltration fibers from NX Filtration BV. Figure 2 As can already be estimated, PAI (>147 MPa) has a higher tensile strength compared to PES (<90 MPa), and the burst pressure results also confirm that PAI fibers exhibit substantially higher burst pressure at similar wall thicknesses. These results highlight the importance of selecting appropriate polymer properties in advance to allow for suitable fibers in FO membrane support design. Appropriate polymer selection becomes particularly crucial when achieving low structural parameters by reducing thickness.

[0135] 1.3.4 Zeta potential of hollow fiber support (before IP) Prior to the application of the IP layer, the zeta potential of all fibers varied with pH, ​​as plotted on... Figure 3 In the figure, the zeta potential of the modified fibers is shown in color for better visualization, while the bare fibers are shown in black. For the modified fibers, a nearly identical trend in zeta potential was observed. They exhibited a more positive zeta potential profile compared to the bare fibers. This difference clearly indicates successful modification of the inner surface with positively charged PEI. However, the zeta potential of the modified fibers does not perfectly reflect the zeta potential expected from PEI alone. To better understand this, the zeta potential of the bare fibers post-modified by PEI (referred to as Post-M) is added to the figure. The difference in zeta potential becomes apparent when comparing the surface charge of Post-M fibers with that of fibers modified in a single step. Post-M fibers exhibit a positive charge across almost the entire pH range, possessing an isoelectric point above 10, consistent with the pKa of PEI. In contrast, the zeta potential of the fibers modified during spinning clearly shows a lower isoelectric point. This difference in zeta potential between fibers modified in a single step and those post-modified suggests that the PAI support also contributes to the final surface charge of the fibers modified in a single step. The significant contribution of the PAI support to the surface charge indicates that the PEI added to the membrane via the pore liquid is integrated into the support due to the single-step preparation.

[0136] 1.3.5 Pure water permeability (PWP) and salt rejection rate The pure water permeability of the fiber support was measured before IP coating. Figure 4 It is noteworthy that a decreasing trend in pure water permeability was observed as the wall thickness of the modified fibers increased. This demonstrates that thickness has a significant impact on mass transfer resistance through the support, as the lowest thickness exhibits the highest water permeability (lowest resistance). In particular, lower mass transfer resistance is beneficial for mitigating ICP occurring within the support during FO. This is because low resistance leads to higher solute mobility within the support and compensates for ICP.

[0137] Compared to modified fibers (M0.12) of similar thickness, the bare fibers showed a higher PWP. Although their porosity differences were small (78.3% for the bare fibers and 76.1% for M0.12), the higher PWP of the bare fibers was primarily attributed to the lack of PEI modification. Due to the added PEI, the M0.12 fibers exhibited significantly additional resistance to water transport, effectively reducing pore size. Clearly, the higher flux of the bare support compared to M0.12 (76.1%) was not due to its higher porosity (78.3%), as there is no clear correlation between porosity and PWP. This further demonstrates the substantial effect of wall thickness on mass transfer within the HF support.

[0138] The NaCl and TSC rejection rates of the IP-coated modified fibers were measured in RO mode, and were respectively... Figure 5A and Figure 5B The graph in Figure A compares the water permeability of the fibers. The first notable observation is that the IP-coated fibers generally exhibit better TSC rejection than NaCl, which can be attributed to the difference in solute size. Secondly, in both cases, the thinner fibers (M0.07) exhibited lower rejection rates, likely due to layer defects caused by their low thickness. The low wall thickness of M0.07 may not provide sufficient reservoir for IP coating, thus lacking enough monomers to participate in the reaction to obtain the defect-free layer used in our IP scheme. This can be overcome by optimizing the IP coating technique to improve rejection rates for such thin hollow fiber supports. Therefore, the lower rejection rate for M0.07 can be attributed to defects in the selective layer. Finally, apart from the M0.07 fibers, the remaining modified fibers exhibited high rejection rates for NaCl and TSC, >88% and >97%, respectively. These high rejection rates demonstrate the high success rate of IP coating and the presence of a defect-free PA layer, a prerequisite for FO membranes.

[0139] It is worth noting that no correlation was found between the wall thickness of the modified fibers after IP coating and the water permeability (A) of the selective layer. This indicates that the properties of the PA selective layer formed on all fibers are independent of their wall thickness. This identical PA selective layer on all modified fibers is actually advantageous, as it will ultimately limit attention to the effect of the support wall thickness on FO performance.

[0140] Furthermore, for all membranes with different wall thicknesses, <1 (L·m) was obtained. -2 h -1 -bar -1 The water permeability (A) of the membrane is considered, and the membrane is placed in an A-limited state. This means that FO performance (permeate flux) is affected by both the selective layer and the support layer. Therefore, it is expected that FO performance can be significantly improved with further optimization of the IP scheme.

[0141] 1.3.6 FO Performance To investigate the effect of wall thickness, IP-coated modified fibers were tested in FO mode. Permeation flux (J) W ) and reverse salt flux (J S ) varies with wall thickness and is drawn in Figure 6A and Figure 6B Our FO membranes achieved permeate flux lower than the typical values ​​reported for FO membranes, primarily due to their low A value. By optimizing the IP coating process, the A value can indeed be increased to greater than 1 (L·m). -2 h -1 -bar -1 The value of A is not the focus of this study. However, although partially influenced by the value of A, the effect of wall thickness on ICP (S-parameter) is already significant, and if the value of A is higher than 1 (L·m), the effect is even greater. -2 h -1 -bar -1 The effect is likely even more pronounced with thinner support (e.g., thinner support, which leads to lower ICP and thus higher permeate flux). It can be seen that reducing the wall thickness increases the permeate flux. This trend is theoretically expected from Equation 1, where a thinner support should result in lower ICP and therefore higher permeate flux. The increasing trend is visible for both draw solutions, with only one exception for M0.07 when using 1M NaCl. This poorer result for M0.07 is due to its lower NaCl rejection (66 ± 3%), which results in higher reverse salt flux during the FO process.

[0142] Besides Jw, the reverse salt flux (Js) also showed an increasing trend when the wall thickness was reduced. Although high Js is detrimental to the FO process due to the loss of driving force, our modified fibers showed acceptable reverse salt flux (except for M0.07). However, this can be addressed by optimizing the IP coating, which will ultimately allow for the appropriate coating of thinner supports.

[0143] Similarly, the measured S-parameters, determined by empirical methods (Equation 8), show the same trend. For both draw solutions, the S-parameters increase with increasing wall thickness. Figure 7A and Figure 7B Generally, a lower S-parameter is obtained when NaCl is used as the draw solution compared to TSC. Haese et al.

[44] found that S increases with increasing valence of the draw solute ions (in our case, NaCl compared to TSC). This finding suggests that the S-parameter of the support depends not only on its physical properties (torsion, porosity, and thickness) but also on other parameters such as the properties of the draw solute. On the other hand, this finding of two different S-parameters for the same support may be due to the inaccuracy of the fitted equation (Equation 8) for estimating ICP. The inaccuracy of the existing structural parameter model (Equation 8) has been discussed previously in [2,45]. Regardless of these challenges, the results presented here clearly demonstrate the importance of the thickness of the support for FO membranes and that supports with the necessary mechanical strength for industrial applications can be fabricated by selecting a suitable polymer support with a very low thickness.

[0144] 1.4 Conclusion In this study, the effect of hollow fiber wall thickness on membrane performance in forward osmosis (FO) mode was systematically investigated. During spinning, the inner surface of the support was simultaneously modified with PEI to facilitate subsequent IP coating. After interfacial polymerization, the membrane exhibited high rejection rates for both NaCl and TSC (typical solutes in the FO process). To ensure the integrity of the membrane for FO applications, the mechanical strength of the hollow fibers is a crucial aspect of membrane design. By selecting PAI, a high tensile strength polymer, we were able to fabricate very thin hollow fiber membranes while maintaining burst pressures well above 10 bar for all membranes. This is more than sufficient for FO applications. We also demonstrate the interrelationships between fiber geometry, polymer tensile strength, and initial fiber burst pressure.

[0145] When 1M TSC and 1M NaCl were used as extractants, by reducing the wall thickness, the permeate flux significantly increased from 1.2 ± 0.4 to 7 ± 1 (L·m⁻¹). -2 h -1) and increased from 2 ± 0.4 to 7 ± 2 (L·m -2 h -1 Even for the thinnest fibers, the relatively low water flux can be attributed to the low water permeability of the membrane after IP (A). It is expected that even higher flux can be obtained by further optimizing IP, and thus increasing water permeability.

[0146] Our results demonstrate that by using suitable polymers and reducing the wall thickness of the support, the permeate flux can be significantly increased without compromising mechanical strength. Therefore, these types of thin hollow fiber supports are promising alternatives to the thick RO-type supports commonly used in FO membranes.

[0147] Example 2 2.1.1 Materials Polyamide-imide (PAI) Torlon® 4000TF was received from Solvay Advanced Polymers and used as the base polymer for coating solution preparation. Branched polyethyleneimine (PEI) with three different molecular weights (PEI 750 k: M) was also used. w ~750,000 g.mol -1 50 wt% solution; PEI 25k:M w ~25,000 g.mol -1 Pure solution; and PEI 2k:M w ~2,000 g·mol -1 A 50 wt% solution was used in pore solution for single-step modification of PAI hollow fiber supports. 1-Methyl-2-pyrrolidone (NMP) 99% was purchased from Acros Organics. Isopar E was supplied amicably by ExxonMobil Petroleum & Chemical B.VBA (Belgium). 1,3-Phenylenediamine (MPD) >98%, ethylene glycol (EG), glycerol 85-87%, and 1,3,5-trimethylbenzene chloride (TMC) >98% were purchased from Sigma-Aldrich. All chemicals were used as accepted without further purification.

[0148] 2.1.2 Manufacturing of hollow fibers The coating composition consisted of 15 wt% PAI, 16 wt% EG, and 69 wt% NMP, and was adapted from the coating formulation used in our Example 1, which comprised less than 17% of the total polymer or a 3 wt% reduction in the PAI content of the coating. The base polymer (PAI) was dried in an oven at 100°C for 24 h prior to coating solution preparation. After all compounds were added to a bottle, both the pore solution and the polymer solution (coating) were prepared on a roller table. The coating was degassed by placing the polymer solution in the coating container of a spinning machine for 24 h. Fibers were spun using a dry-jet wet phase inversion spinning process.

[0149] Torlon (PAI) was chosen for its excellent mechanical properties and the potential for interaction with PEI, which respectively result in high burst pressure of the fiber and proper adhesion of PEI to the support. We modified the PEI to facilitate subsequent interfacial polymerization by adding amines to the inner surface of the support.

[0150] To investigate the success of PEI modification, solutions with pores containing PEI of different molecular weights were used, as well as solutions with pores without PEI. In total, four different types of fibers were spun in this study (Table 4). The first batch (unmodified) was fabricated using pore B1 as a reference for further analysis, while the other three fibers were fabricated using pores containing one of the different PEIs (B2, B3, or B4).

[0151] Table 4. Pore compositions used in the spinning process described in this work.

[0152]

[0153] The initial fibers were then rinsed in deionized (DI) water for 72 h, followed by post-treatment by immersing them in a 30% w / w glycerol solution for 24 h to preserve pores during drying. Finally, the fibers were hung to dry at room temperature until further use.

[0154] Table 5. Spinning conditions for manufacturing four types of initial fibers.

[0155]

[0156] The newly spun fiber (fiber 750K) was compared with a fiber (i.e., M0.12) spun with 3 wt% more polymer in its coating formulation and whose surface was modified with 5 wt% PEI 750k.

[0157] 2.1.3 Component Preparation Five fibers were placed inside a 10 mm perfluoroalkoxy (PFA) tube, with two QST-10 Festo push-in fittings attached to each side. Both ends were sealed with a 4.5 cm length of slow-curing resin (BISON, 2K Expert Polyurethane) to achieve a fiber length of 16 cm. Three to four assemblies were prepared for further experiments for each fiber type. All filtration experiments were run in at least triplicate using individual assemblies.

[0158] 2.2 Membrane Characterization 2.2.1 Scanning Electron Microscopy Hollow fibers were frozen and fractured in liquid nitrogen to prepare cross-sectional samples for imaging. To prepare inner surface samples, the fibers were diagonally cut in half with a sharp blade. All samples were mounted on designated holders and stored in a vacuum oven at 30°C for 24 h. Finally, before microscopic examination using a JEOL JSM-6010LA SEM, the samples were sputter-coated with a 5 nm platinum layer using a Quorum Q150T ES coating machine.

[0159] 2.2.2 Contact Angle Measurement The water contact angle of the inner surface was measured to investigate the effect of PEI modification on the hydrophilicity of the support surface. For this purpose, the hollow fiber support was cut and mounted on double-sided tape so that the inner surface could be exposed to the analyzer's microsyringe. Figure 1 A 0.5 µL droplet of Di water was placed on the inner surface. The contact angle of the droplet was recorded after 3 seconds and analyzed using software. The measurement was repeated three times at three different points for each support.

[0160] 2.2.3 Mechanical Strength The burst pressure of the single fiber was determined using a custom setup manufactured by Convergence Industry BV, Netherlands. It forces fluid through the hollow fiber under increasing pressure until the sample fails. This ultimate pressure was reported. For this purpose, a single-fiber membrane module was fabricated using a material similar to that used in the filtration experiments. Pressure was applied from the inside of the membrane in increments of 0.5 bar per second. When using the fluid glycerol, less permeation is allowed due to its high viscosity, and therefore pressure control is easier. The experiments were conducted at room temperature.

[0161] Under ultimate pressure, the tensile strength (δ) of the fiber is determined by Barlow's formula, given in Equation 1.

[0162] (1) Where P is the obtained burst pressure (bar), t is the wall thickness (mm), and D... 外δ is the outer diameter (mm) and δ is the tensile strength (bar).

[0163] 2.2.4 Porosity Measurement Before testing, the fibers were rinsed for 20 min in an inside-out configuration and kept in a vacuum oven at 30°C for at least 4 h. Porosity was measured using the known value of the polyamide-imide density. The sample was placed in a vacuum oven for at least 1 hour before its weight was measured. Porosity was calculated using Equation 2.

[0164] (2) m ρ is the sample mass (g), ρ is the PAI density (g·cm⁻³), and V is the fiber volume (cm³). -3 The density of the polyamide-imide used in this calculation is 1.4 g·cm³. -3 Each type of fiber was tested in triplicate.

[0165] 2.2.5 Determination of ζ potential The zeta potential of the fibers was determined at pH (3-11) using a SurPASS electric analyzer (Anton Paar, Graz, Austria) as the electrolyte solution, with 5 mM KCl as the electrolyte solution as the instrument. For this purpose, two single fibers of each type were encapsulated in separate 7 cm long tubes. The zeta potential was determined by allowing the electrolyte solution to flow within the fiber lumen, with six measurements taken for each pH setting for each fiber. The zeta potential value was then determined by averaging all 12 measurements at each pH. The pH was adjusted by automatic titration using 0.1 M HCl and 0.1 M NaOH. The zeta potential was calculated using Equation 3.

[0166] (3) In this equation, ζ is the zeta potential (V), I is the flowing current (A), P is the pressure (Pa), η is the dynamic viscosity of the electrolyte solution (Pa·s), ε is the dielectric permittivity of water, and ε0 is the dielectric permittivity in vacuum (F·m). -1 ), L s It is the channel length (m), and A s It is the cross-section of the flow channel (m) 2 ).

[0167] 2.2.6 Measurement of Pure Water Permeability The pure water permeability (PWP) of the fiber was measured using compressed nitrogen at 1 bar in a dead-end, inside-out mode. (A) (L·m) -2 h -1 -bar -1).

[0168] (4) Where V is the volume of permeate collected (L), t is the collection time (h), and A is the membrane inner surface area (m²). 2 ), and ΔP is the pressure difference (bar) between the inner cavity and the shell of the hollow fiber.

[0169] 2.2.7 Interface Aggregation A polyamide (PA) selective layer was formed by coating the inner surface of fibers using interfacial polymerization of MPD and TMC. Prior to the reaction, all fiber assemblies were run at 1 bar in a dead-end, inside-out mode in a PWP test to remove glycerol from the fibers. A 2.5% (w / v) MPD solution was prepared by dissolving the required amount of MPD in DI water. A 0.15% (w / v) TMC solution was obtained by dissolving TMC in Isopar E using an ultrasonic bath (20°C, 2 min).

[0170] First, excess water in the inner cavity and on the shell side of the hollow fiber is removed by passing N2 through a flow for 1.5 min. Then, MPD solution is injected into the inner cavity side of the vertically fixed fiber at 10 L·min⁻¹ using a syringe pump for 2 min. The solution is then held horizontally within the fiber for 10 min, followed by removal of excess MPD by passing N2 through a flow for 1 min. Subsequently, the solution is injected at 10 L·min⁻¹ using a syringe pump. -1 The TMC solution was fed into the vertically fixed fibers twice for 15 seconds, with a 45-second interval. Residual TMC was then removed by passing it through a nitrogen stream for 30 seconds, followed by drying in an air oven at 70°C for 10 minutes. Finally, the fiber cavity was rinsed with 20 ml of DI water, and the assembly was kept in DI water at 7°C for further use.

[0171] 2.2.8 Salt Retention Rate For 500 mg·L -1 The fiber rejection rate of NaCl solution was determined in a cross-flow filtration system at a transmembrane pressure (TMP) of 2 bar. The rejection rate was calculated using Equation 5 by measuring the conductivity of the feed solution and the collected permeate. The cross-flow velocity per fiber was set to 1.7 m / s. -1 .

[0172] (5) In this equation, R is the salt rejection rate (%), and σ is the salt rejection rate (%). f The conductivity of the feed solution (μS·cm) -1 ), and σ p It is the conductivity of the permeate solution (μS·cm)-1 ).

[0173] 2.2.9 Permeability The fiber permeability was evaluated in both FO (selective layer-to-layer feed solution) and PRO (selective layer-to-layer draw solution) modes. Tests were conducted in a countercurrent configuration using 1 M NaCl as the draw solution, with a cross-flow rate of 3 L / h per component. -1 (corresponding to 2 m·s in the inner cavity) -1 The experiment was conducted for 1 hour, and the permeate flux (Jw) (L·m) was determined by measuring the amount of water transferred from the feed solution to the draw solution per membrane area each time. -2 h -1 Reverse salt flux (Js) (g·m⁻²) h -1 The conductivity of the feed solution per membrane area is determined by measuring the conductivity each time. The amount of solute passing through the feed side is measured using a pre-defined calibration curve.

[0174] The structural parameters (S) were evaluated using a fitting method; for the FO mode, Equation 6 was used. (6) And for PRO mode, use Equation 7. (7) J w (L·m -2 h -1 D (m) is the infiltration flow rate. 2 ·h -1 π is the solute diffusivity. D (Bar) is used to draw solution A (water-permeable) (L·m -2 h -1 -bar -1 ) and B (salt permeability) (L·m -2 h -1 ) osmotic pressure.

[0175] Based on the Lévêque interrelationship for the fully developed laminar velocity profile in the pipe

[37] , and according to Yip et al.

[36] using the mass transfer coefficient (K) (L·m⁻²) h-1) determines salt permeability (B).

[0176] (8) 2.3 Results and Discussion 2.3.1 Inherent characteristics of hollow fiber supports The inherent characteristics of the manufactured hollow fiber supports, such as their dimensions, porosity, burst pressure, and contact angle, can be found in Table 6. To better understand the effect of the molecular weight of the PEI used, an attempt was made to eliminate the influence of wall thickness on FO performance by manufacturing fibers with similar thicknesses. This is because wall thickness affects permeate flux (see Example 1) and is therefore difficult to focus on the effect of molecular weight. With relatively low wall thickness and open-cell structure, the spun fibers still exhibited a high burst pressure of approximately 14 bar on average. This is indeed due to the high tensile strength of the polyamide-imide (Torlon). We believe that a burst pressure of 10 bar would be sufficient for membranes used in FO, similar to the burst pressure of commercial ultrafiltration membranes, where operating pressures are typically below 1 bar. Therefore, all membranes are mechanically suitable for the FO process. Variations in PEI molecular weight showed no effect on burst pressure and no difference from the unmodified support. The fibers also exhibited a porosity of >73%, indicating that they are very open and suitable as FO supports. In fact, for porous media, a porosity greater than 75% also means low tortuosity, both of which are beneficial for low ICP and therefore for high permeable water flux, see Example 1.

[0177] Here, a comparison is made with the fiber of Example 1, which was spun with an additional 3 wt% polymer (i.e., M0.12), to examine the effect of a 3 wt% reduction in polymer content on the inherent properties of the fiber. We demonstrate throughout these examples that a 3 wt% reduction in polymer content is reflected very well in the fiber's PWP, permeate flux, S-parameters, and mechanical strength.

[0178] For example, a 3 wt% reduction in polymer is well reflected in the PWP results, with the newly spun 750K achieving a PWP seven times higher. This higher PWP clearly indicates a more open support structure. Furthermore, the 3 wt% reduction also affected the burst pressure, decreasing it from 18 bar to 14 bar, which, as previously discussed, is still sufficient for FO applications.

[0179] Table 6. Dimensions, porosity, and burst pressure of the spun fibers measured, n=3, dimensions determined by the SEM dimensioning tool. (M0.12 was spun with an additional 3 wt% polymer content; data taken from Example 1 for comparison).

[0180]

[0181] All three modified supports (2K, 25K, 750K) exhibited lower contact angles than the unmodified support, indicating that PEI contributes to the hydrophilicity of the inner surface. Indeed, the hydrophilic surface promotes wettability by allowing the IP (amine) solution to penetrate all surface pores, and thus favors interfacial polymerization. This ensures deeper formation of the polyamide layer within the support pores, and therefore better adhesion of the initial PA layer. Overall, variations in PEI molecular weight did not show significant differences in the intrinsic properties of the support.

[0182] 2.3.2 Morphology of hollow fibers SEM images of unmodified fibers and 750K fibers in Figure 8 The images are shown for comparative purposes. The morphology of the combined finger-like and sponge-like structures is observed in cross-sectional images of the two fibers (unmodified -(a) and 750K-(a) images). This similarity shows that PEI surface modification has no effect on the overall fiber structure, which is actually advantageous because it demonstrates that the overall structure of the support can be independently adjusted while maintaining a constant inner surface. Both types of fibers have an open structure, highly open in the middle of the fiber, which is beneficial in terms of reducing ICP (unmodified -(b) and 750K-(b) images). A fairly visible difference in the inner surface of the unmodified and 750K fibers can be seen by observing the images of unmodified -(c) and 750K-(c). The unmodified fiber exhibits a rough, uneven, and defective inner surface, while the modified fiber (750K) has a very smooth, uniform, and defect-free inner surface. Similar surface structures are seen for the other two HF support types modified with PEI (2K and 25K fibers). Furthermore, these similarities indicate that the molecular weight of PEI has little effect on the surface structure. It is clear that by using PEI, defective and rough inner surfaces can be avoided. Defective surfaces are detrimental because they lead to imperfections (pinholes) on the selective layer, resulting in unsuccessful IP coating. Image (d) shows the inner surfaces of both the untreated and 750K surfaces after IP application. Figure 8 In d 750K (right), a very typical and uniform PA layer with a ridge-valley structure can be observed, which is not the case for the unmodified fiber. The same PA layer structure is seen for the other two fiber types modified with PEI (2K and 25K fibers – not shown here). Our results vividly demonstrate that PA fabrication can be significantly improved by modifying the support with PEI, regardless of its molecular weight.

[0183] Furthermore, the cross-sectional morphology of fibers 750K and M0.12 is in Figure 9The image shows the effect of a 3 wt% polymer reduction on morphology. 750K fibers exhibit a finger-like structure extending from the lumen to the middle of the fiber, combined with a highly open structure extending to the shell. In contrast, M0.12 fibers have finger-like structures from both sides, with a relatively denser middle portion. The more open structure of 750K compared to M0.12 is evident in this cross-sectional image. We believe that the denser internal portion of the M0.12 fibers, in particular, significantly contributes to ICP during the FO process.

[0184] 2.3.3 Zeta potential of hollow fibers before IP The ζ-potential at the inner surface of the HF support was determined. Figure 10 This was done to evaluate surface modification with PEI. Typically, the three modified fibers showed more positive zeta potentials than their unmodified counterparts, confirming the presence of positively charged PEI. This result is consistent with our previous study, where PEI with a molecular weight of 750 kJ was used for single-step modification. Figure 10 The addition of M0.12 to the zeta potential indicates that the 3 wt% polymer reduction shows no effect on the PEI modification of the surface. However, the zeta potentials of all three fibers do not fully reflect the zeta potential of pure PEI (as explained in Example 1). Lower zeta potentials were observed within the studied pH range, suggesting that the support base polymer (unmodified PAI) influences the final zeta potential. This implies that PEI has become an integral part of the support and therefore has proper adhesion to the entire support.

[0185] 2.3.4 Pure water permeability (PWP) and salt rejection rate First, the pure water permeability of the fiber was measured before IP coating. Figure 11 All four fiber types exhibited high purity water permeability, indicating a very open structure. This is advantageous because the open structure implies low mass transfer resistance, which facilitates the use of the support in FO applications. Low transport resistance leads to rapid solute transport to the selective layer, which in turn reduces ICP. Modified HF supports (2K, 25K, and 750K) showed slightly lower permeability compared to unmodified fibers. This suggests that single-step modification results in only a slight resistance to mass transfer; therefore, a small effect on ICP is expected. The 750K fiber showed a significant increase in PWP compared to its counterpart (M0.12) spun with a 3% higher polymer concentration. Figure 11 For 750K, this is as high as 365 ± 64 (L·m). -2 h -1 -bar -1 ), and for M0.12, it is 48 ± 2 (L·m -2 h-1 -bar -1 The study showed that the structure of the newly spun fiber (750K) was significantly more open due to a 3% reduction in polymer content.

[0186] The PWP of all four fiber types decreased significantly to approximately 1 (L·m) after IP treatment. -2 h -1 -bar -1 ()( Figure 12 (x-axis). We observed no significant effect on the molecular weight of PEI used to prepare the modified membrane. The unmodified +IP HF membrane (square) showed higher water permeability and greater variation than the modified fiber. This indicates that the IP process of the unmodified +IP membrane was unsuccessful because it lacked PEI on its surface, which also resulted in a low NaCl rejection rate (x-axis). Figure 12 (Y-axis). Conversely, all three modified HF films showed high success rates for IP, with NaCl rejection rates exceeding 77% and low data bias. In particular, the results for M0.12 from Example 1 are added to the figure, which also shows low bias in the data as PEI results on its surface. However, M0.12 showed a lower PWP than 750 K, which is partly attributed to its denser support due to the 3 wt% more polymer in its structure.

[0187] 2.3.5 Permeability The permeation properties of IP-coated fibers were determined in both FO and PRO modes using 1M NaCl as the draw solution and DI water as the feed solution. Figure 13 ).like Figure 13 As shown, all four types of membranes (unmodified +IP, 2K+IP, 25K+IP, 750K+IP) exhibited values ​​higher than 9 (L·m). -2 h -1 The similarity of the permeate flux of the four membranes clearly confirms that when A (water permeability) is higher than 1 (L·m⁻¹), the permeate flux is significantly higher. -2 h -1 -bar -1 When the permeate flux is primarily influenced by the support structure (properties) of the FO membrane, and to a lesser extent by its selective layer, our membrane was obtained. All four types of fibers have approximately the same structure, porosity, and wall thickness; therefore, it is assumed that the ICP levels are similar, leading to roughly the same water flux. However, defects in the selective layer of the unmodified +IP membrane resulted in a high reverse salt flux (>34 g·m³). -2 h-1 On the other hand, the modified fibers exhibit low reverse salt flux (<2.5 g·m). -2 h -1 The presence of a defect-free PA layer is attributed to the PEI surface modification. The membrane demonstrated a water flux of 12 ± 2 (L·m) at 750 K+IP. -2 h -1 This is achieved by spinning its counterpart (M0.12) with 3% more polymer content, which is 4 times more efficient. This clearly demonstrates that the more open support structure of 750+ IP is beneficial for achieving permeable water flux.

[0188] In PRO mode ( Figure 14 In this study, all four types of fibers spun exhibited water fluxes higher than 15 (L·m). -2 h -1 The permeate flux in PRO mode is higher than that obtained in FO mode, which is typical for permeation processes. This is because the concentrated ICP that occurs in PRO mode is much less harmful than the diluted ICP that occurs in FO mode. The comparison of permeate flux between 750K+IP and M0.12+IP in PRO mode follows the same trend as in FO mode, where a higher flux is achieved at 750K+IP. The reverse salt flux of the membrane presented here in PRO mode is comparable to that in FO mode.

[0189] All membrane structural parameters were calculated using equations (6) and (7) and plotted on [the graph]. Figure 15 In the FO mode, the S-parameter of the modified membrane was determined to be approximately 0.6 mm, while the S-parameter of the unmodified +IP membrane was higher (0.75 mm). This can be attributed to the leakage characteristics of the selective layer, which is not properly explained in Equations 6 and 7. Notably, the structural parameter of M0.12 is as high as 1.7 ± 0.4 mm, compared to 0.55 ± 0.05 mm for the newly spun membrane (750K +IP). In the PRO mode, the modified membrane exhibited a higher S-parameter compared to the FO mode, while the S-parameter of the unmodified +IP membrane was lower. This difference suggests that the sensitivity of the S-parameter is a true characterization of the support.

[0190] 2.4 Conclusion This study presents a single-step method for fabricating open hollow fiber membrane supports (HF) modified with PEI. Surface modification with PEI was performed to produce a smooth and defect-free inner surface for more successful IP coating. Surface modification was carried out simultaneously during the spinning of the support using PEI with three different molecular weights. To fabricate open fibers, we used 17% less polymer in our polymer coating solution, a 3 wt% reduction in the composition of the coating compared to our previous study. Surprisingly, the reduction in polymer concentration was not reflected in the porosity of the newly spun fibers; however, this became apparent through SEM images, PWP, permeate flux, and S-parameters. The latter three all showed significant increases. The initial HF support showed a porosity of approximately 400 (L·m⁻¹). -2 h -1 -bar -1 The high PWP of the modified fibers indicates that they are highly open, which can promote low ICP. After IP application, the modified fibers showed a sufficient NaCl rejection rate of >77% compared to the average of 27% for unmodified membranes, demonstrating that surface modification provides a defect-free surface. In FO mode, the average permeate flux exceeded 9 (L·m). -2 h -1 In PRO mode, this exceeds 15 (L·m) -2 h -1 The modified fibers exhibited very low reverse salt flux in both modes. The most significant effect of a 3% reduction in polymer content was a 4-fold increase in FO water flux and a significant decrease in S-parameters. Furthermore, changes in PEI molecular weight had no effect on initial HF support properties such as burst pressure, water flux, internal surface morphology, NaCl rejection, and permeability.

[0191] This study demonstrates that PEI surface modification significantly improves the IP reaction, regardless of molecular weight. Furthermore, it proves that highly open supports with improved permeability can be fabricated through directional design, which is advantageous for use in FO membranes. This novel approach and the chosen polymer will enable the fabrication of very thin, open, defect-free, and robust HF supports for FO membranes.

[0192] Various aspects and implementations have been described in conjunction with multiple embodiments herein. However, by studying the accompanying drawings, disclosure, and appended claims, those skilled in the art will understand and implement other variations of the disclosed embodiments in practicing the claimed subject matter. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite articles "a" or "an" do not exclude a plurality. A single processor or other unit can perform the functions of multiple items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not imply that a combination of these measures cannot be used advantageously. Computer programs may be stored / distributed on suitable media, such as optical storage media or solid-state media provided with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunications systems.

Claims

1. A method for preparing hollow fibers, comprising the following steps: a) Provide coating solutions comprising polyamide-imide (PAI), b) Provide an aqueous porous solution comprising polyethyleneimine (PEI), c) Extruding the coating solution in a cross-sectional annular shape and spraying the aqueous pore solution at the center of the annular shape. d) Allowing the PAI and PEI to react, thereby forming an inner surface layer comprising the crosslinking reaction product, and e) A polyamide film composite TFC layer is formed by an interfacial reaction of an aqueous diamine or triamine compound and an organic diacyl halide or triacyl halide compound on the inner surface layer comprising the crosslinking reaction product, such that the inner surface layer is used as a primer to connect the TFC layer to the PAI support layer.

2. The method according to claim 1, wherein, The monomer of PAI has the following chemical structure: Ar is selected from the following aromatic groups: , , or .

3. The method according to claim 1, wherein, The coating solution further includes an aprotic solvent.

4. The method according to claim 3, wherein, The aprotic solvent is selected from the group consisting of: N-methylpyrrolidone (NMP), dimethylacetamide (DMAC), dimethyl sulfoxide (DMSO), and dimethylformamide (DMF), or mixtures thereof.

5. The method according to any one of claims 1 to 4, wherein, The coating solution further comprises polymers selected from the group consisting of: polyphenylene sulfone (PPSU), polyether sulfone (PES), sulfonated polyether sulfone (SPES), polysulfone (PSU), polyether ether ketone (PEEK), polybenzimidazole (PBI), polyimide (PI), polyether imide (PEI), polyphenylene sulfide (PPS), and mixtures thereof.

6. The method according to any one of claims 1 to 4, wherein, The coating solution further comprises polymers selected from the group consisting of: high-temperature sulfone resins, self-reinforced polystyrene and mixtures thereof.

7. The method according to any one of claims 1 to 4, wherein, The PAI solution, measured in 25% NMP at 40ºC, has a viscosity of 30,000 cP or less.

8. A hollow fiber, comprising an outer support layer, an inner surface layer, and a polyamide film composite TFC layer, wherein, The outer support layer comprises polyamide-imide (PAI), the inner surface layer comprises a reaction product of PAI and polyethyleneimine (PEI) that is in direct contact with the TFC layer, and the TFC layer comprises a reaction product of a diamine or triamine compound and a diacyl halide or triacyl halide compound.

9. The hollow fiber according to claim 8, wherein, The outer support layer further comprises polymers selected from the group consisting of: polyphenylene sulfone (PPSU), polyether sulfone (PES), polysulfone (PSU), polyether ether ketone (PEEK), polybenzimidazole (PBI), polyimide (PI), polyether imide (PEI), polyphenylene sulfide (PPS), and mixtures thereof.

10. A hollow fiber assembly comprising a bundle of hollow fibers according to any one of claims 8 to 9.

Images