Porous membrane and liquid filtration method using the porous membrane

Abstract

To provide a porous film having high dirt resistance and highly accurate removal property.SOLUTION: Provided is a porous film in which, on at least one surface, a surface part from the surface to a thickness of 10 μm is denser than the inside part, a removal rate T of dextran having a weight average molecular weight 40,000 Da is 60 to 95%, and the number of surface holes per unit area observed on the surface of the surface part is 200 / μm2 to 2,000 / μm2.SELECTED DRAWING: None

Description

[Technical Field]

[0001] The present invention relates to a porous membrane and a method for producing the porous membrane. [Background technology]

[0002] In recent years, porous membranes such as microfiltration membranes and ultrafiltration membranes have been used in a variety of fields, including water treatment (such as water purification or wastewater treatment), medical treatment (such as blood purification), and the food industry. Recently, there has been a demand for highly fouling-resistant porous membranes that can efficiently filter "filtrate stock solutions that are prone to fouling," which have been considered difficult to filter. Filtrate stock solutions that are prone to fouling porous membranes are characterized by containing more fouling substances than conventional stock solutions. These fouling substances often contain coarse components (Stokes diameter: 13 nm or greater) that are the target of removal, and fine components (molecular weight: 10,000 Da or less) that are not the target of removal but are preferably permeable as useful substances, etc.

[0003] The fouling resistance of a porous membrane means that when performing high-precision removal, problems such as clogging of the pores of the porous membrane can be reduced for both coarse and fine components that are fouling substances or the objects to be removed, and the removal efficiency and water permeability of the objects to be removed can be maintained stably for a long period of time.

[0004] As a porous membrane that exhibits fouling resistance, Patent Document 1 discloses a porous membrane that has a three-dimensional mesh structure with large pores, thereby achieving both a high removal rate of the target substance and a high water permeability coefficient.

[0005] Patent Document 2 also discloses a technology for improving the removal rate of target substances such as viruses by miniaturizing the pore size of a three-dimensional mesh structure. The three-dimensional mesh structure disclosed in Patent Document 2 has an average diameter of 0.01 μm or more and 1 μm or less, and is substantially free of macrovoids of 5 μm or more. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2006-82006 [Patent Document 2] Japanese Patent Publication No. 2010-94670 Summary of the Invention [Problem to be solved by the invention]

[0007] However, the three-dimensional network structure with large pores described in Patent Document 1 has a large average pore size and a spider web-like three-dimensional structure, allowing coarse components to penetrate and foul the porous membrane. Furthermore, since fine particles are trapped in the gaps between the tangled microfilament structures, the porous membrane is prone to clogging and has insufficient fouling resistance. The porous membrane described in Patent Document 2 only targets relatively large components such as viruses as targets for removal, improving removal performance. The network structure has a relatively large pore size and a small number of pores. Therefore, fouling substances tend to accumulate within the porous membrane, sacrificing fouling resistance when used with easily fouling filtrate stock solutions.

[0008] On the other hand, since the porous membrane is easily soiled with the above-mentioned raw solution, filtration is difficult, and there is a demand for a porous membrane with high fouling resistance that enables filtration even with such easily soiled raw solution.

[0009] Therefore, an object of the present invention is to provide a porous membrane having excellent fouling resistance by removing coarse components at the outer surface of the porous membrane and allowing fine components to efficiently permeate the porous membrane. The outer surface is the surface portion of the porous membrane, in other words, the outer surface of the surface layer opposite to the inner surface. [Means for solving the problem]

[0010] In order to solve the above problems, the present invention provides a porous membrane having the following configuration. 1. On at least one surface, the surface portion from the surface to a thickness of 10 μm is denser than the interior, the dextran removal rate (T) of which is 60% to 95% and the number of pores observed on the surface of the surface portion (hereinafter referred to as surface pores) per unit area is 200 / μm 2 ~2000 pieces / μm 2 Yes The average value of the surface pore diameter [nm] of the surface portion is 5.0 nm to 12 nm, and the number of pores observed in a cross section of the outermost surface portion (hereinafter referred to as cross-sectional pores) per unit area is 100 / μm in the outermost surface portion from the surface to a thickness of 2 μm. 2 ~1000 pieces / μm 2 and has a nano-mesh structure with an average cross-sectional pore diameter [nm] of 1 nm to 99 nm. 2 The number of surface pores per unit area [number / μm 2 ] by the average surface pore diameter [nm] of the surface portion: X is 30 to 100 pores / μm 2 / nm 1 to The porous membrane described is 3 The number of cross-sectional holes per unit area [number / μm 2 ] by the average cross-sectional pore diameter [nm] of the outermost surface portion: Y is 3 to 10 pores / μm 2 / nm 1 or 2 It is a porous membrane described in 4 The standard deviation of the cross-sectional pore diameter [nm] of the outermost surface portion is 1.0 nm to 50 nm. 1~3 It is a porous membrane described in 5 The standard deviation of the surface pore diameter [nm] of the surface portion is 0.5 nm to 5.0 nm. 1~4 It is a porous membrane described in

[0011] 6 .1~ 5 A method for filtering a liquid using the porous membrane according to any one of the above. [Effects of the Invention]

[0012] According to the present invention, a porous membrane with excellent fouling resistance can be provided by removing coarse components on the outer surface of the porous membrane and efficiently allowing fine components to permeate through the porous membrane. This allows for highly accurate filtration of a filtrate that is prone to fouling, while suppressing clogging due to fouling and maintaining a high filtration flux even during long-term use. [Brief explanation of the drawings]

[0013] [Figure 1] FIG. 1 is a schematic diagram showing the filtration state using the porous membrane of the present invention. [Figure 2] FIG. 2 is a schematic diagram showing the filtration state using a conventional porous membrane. [Figure 3] FIG. 3 is a schematic diagram showing the filtration state using a conventional porous membrane. [Figure 4] FIG. 4 is a schematic diagram of the process for forming the porous membrane of the present invention. [Figure 5] FIG. 5 is an electron microscope photograph showing a cross section (part) of a porous membrane according to one embodiment of the present invention. [Figure 6] FIG. 6 is an electron microscope photograph showing a cross section (part) of a conventional porous membrane. [Figure 7] FIG. 7 is a characteristic diagram showing the relationship between the number of surface pores and the dextran removal rate. [Figure 8] FIG. 8 is a characteristic diagram showing the relationship between the number of surface pores and the average surface pore diameter. [Figure 9] FIG. 9 is a characteristic diagram showing the relationship between the initial ratio of filtration flux and surface characteristics. [Figure 10] FIG. 10 is a characteristic diagram showing the relationship between the initial ratio of filtration flux and the self-diffusion coefficient. DETAILED DESCRIPTION OF THE INVENTION

[0014] The following describes embodiments of the present invention in detail with reference to the drawings, but the present invention is not limited thereto. In this specification, "mass" and "weight" are synonymous. The term "filtrate stock" refers to the liquid to be filtered before it passes through a porous membrane, and the term "permeate" refers to the liquid after it passes through the porous membrane. A filtrate stock that is difficult to filter refers to a liquid containing a large amount of contaminants that can clog the porous membrane. Contaminants include turbidity, fungi, viruses, polysaccharides, proteins, organic compounds, and complexes thereof. Contaminants often contain coarse components (Stokes diameter: 13 nm or greater) that are the target of removal, and fine components (molecular weight: 10,000 Da or less) that are preferably permeable, such as useful substances. Although fine components should pass through the porous membrane, their small size makes them susceptible to capture by the porous membrane, leading to clogging. The porous membrane of the present invention exhibits high fouling resistance even for filtrate stock containing a large amount of both coarse and fine components.

[0015] With regard to fouling resistance, fouling refers to the clogging of the porous membrane by coarse components (Stokes diameter: 13 nm or more) or fine components (molecular weight: 10,000 Da or less) contained in the raw filtrate during the filtration process. As the porous membrane becomes increasingly clogged with coarse and fine components, the filtration resistance increases and the amount of liquid that can pass through the porous membrane decreases.

[0016] Generally, in order to remove coarse components of turbidity accumulated in a porous membrane from the pores, a fluid such as permeate or compressed air is passed from the direction of the permeate toward the raw filtrate, i.e., backpressure washing is performed in the opposite direction to filtration to remove the coarse components trapped in the porous membrane. Backpressure washing makes it easy to remove coarse components that tend to accumulate outside the porous membrane, but it is difficult to remove fine components that tend to accumulate inside the porous membrane. The improved "fouling resistance" of the present invention makes the pores of the porous membrane less likely to clog, even if the raw filtrate contains both coarse and fine components. The porous membrane of the present invention with improved fouling resistance is described below.

[0017] <About porous membranes> In the porous membrane according to an embodiment of the present invention, at least one surface of the porous membrane is denser than the interior in a thickness of 10 μm from the surface, and the dextran removal rate (T) of the dextran having a weight-average molecular weight of 40,000 Da is 60 to 95%, and the number of surface pores in the surface portion is 200 / μm. 2 ~2000 pieces / μm 2 The surface pores are pores observed on the surface of the surface portion.

[0018] FIG. 1 is a schematic diagram showing the filtration of a difficult-to-filter raw solution using the porous membrane of the present invention. FIG. 1 is a conceptual schematic diagram showing a portion of the cross section of a porous membrane. Porous membrane 101 includes a surface region 102 and an interior region 103. The surface region 102 has a denser structure than the interior region 103. In FIG. 1, the solid line drawn on the surface region 102 schematically shows the mesh structure. The filtration direction FL is the direction from the surface region 102 toward the interior region 103. The filtration target of the raw solution contains coarse contaminant components 201 and fine contaminant components 202. The coarse contaminant components 201 cannot penetrate into the porous membrane because the micropores in the surface region 102 are dense, and do not clog the surface region 102 of the porous membrane, making the porous membrane less susceptible to fouling. On the other hand, since there are many pores in the surface portion 102, minute dirt components 202 are dispersed and penetrate from the surface portion 102 to the interior 103 of the porous membrane in the filtration direction, making the porous membrane less susceptible to dirt.

[0019] In other words, the surface region from the surface to a depth of 10 μm is denser than the interior, preventing coarse contaminants in the filtrate and other target substances from penetrating the porous membrane, demonstrating high fouling resistance. The surface region being denser than the interior can be confirmed by observing the cross section of the porous membrane at 10 μm-thick regions with a scanning electron microscope (SEM) and finding that the porosity of the surface region from the surface to a depth of 10 μm is lower than the porosity of the interior region 10 μm or thicker from the surface. To calculate the porosity, the SEM image of the cross section of the porous membrane is binarized using the free software "ImageJ." To binarize, create a background with a 1-pixel subtraction, then select "Persentile" as the threshold. In the resulting binarized image, select "Area" under "Analyze Particles" to determine the area of ​​the pores. This is then divided by the area of ​​the observed porous membrane to calculate the porosity. The porosity of the surface portion should be smaller than that of the interior, but is preferably 55 to 90%, and more preferably 63 to 80%.

[0020] The surface portion being denser than the interior is preferable because it allows coarse contaminants and substances to be removed that are retained on the outside of the surface portion to be efficiently discharged from the porous membrane, particularly when a backpressure cleaning step in which filtration is performed in the reverse direction during filtration is performed.It is also preferable when crossflow filtration in which liquid is flowed perpendicular to the filtration direction or a flushing step is performed during filtration, because it allows coarse contaminants and substances to be removed that are retained on the surface portion to be efficiently discharged from the porous membrane.

[0021] The inventors discovered that in order to prevent the penetration of coarse components (Stokes diameter: 13 nm or greater) into the porous membrane and to efficiently remove small amounts of coarse components that have penetrated the porous membrane by back pressure washing or the like, it is necessary for the dextran removal rate (T) to be 60% or greater for a dextran with a weight-average molecular weight of 40,000 Da. By specifying the removal rate of dextran with a specific molecular weight, it is possible to obtain a porous membrane with excellent fouling resistance that prevents fouling substances from penetrating the surface and allows fouling substances to be efficiently removed from the porous membrane by back pressure washing or the like.

[0022] The dextran rejection rate (T) is preferably 68% or higher, and more preferably 70% or higher. It has been found that a dextran rejection rate (T) of 95% or lower allows fine components (molecular weight: 10,000 Da or less), which are not targeted for removal but are preferred to permeate, to permeate sufficiently without being trapped in the porous membrane, thereby exhibiting high fouling resistance. A dextran rejection rate (T) of 95% or lower reduces the permeation resistance of the porous membrane, resulting in a high filtrate volume. A dextran rejection rate (T) of 90% or lower is more preferable, and 85% or lower is particularly preferable. A dextran rejection rate (T) of 60 to 95% for dextran with a weight-average molecular weight of 40,000 Da sufficiently prevents coarse fouling substances and substances targeted for removal in the raw filtrate from penetrating the membrane, thereby exhibiting excellent fouling resistance. A dextran rejection rate (T) of 68 to 90% is preferable, 70 to 90% is more preferable, and 70 to 85% is most preferable.

[0023] The dextran removal rate (T) can be calculated using the following formula (3): a dextran aqueous solution prepared at 25°C and containing 1,000 ppm of commercially available dextran with a weight-average molecular weight of 40,000 Da is filtered through a porous membrane at a cross-flow linear velocity of 1.0 m / sec and a transmembrane pressure difference of 10 kPa. T = {(Refractive index of the undiluted solution) - (Refractive index of the transmitted solution)} / (Refractive index of the undiluted solution) × 100 Equation (3) Here, the cross-flow linear velocity is the flow rate of the raw solution in a direction perpendicular to the filtration direction divided by the cross-sectional area of ​​the flow path, and the transmembrane pressure is the difference between the pressure on the raw solution side and the pressure on the permeate side across the porous membrane.

[0024] The number of surface pores observed on the surface of the porous membrane is 200 / μm 2 ~2000 pieces / μm 2 By having a surface pore size of 200 / μm, the contaminants in the raw filtrate can be dispersed within the porous membrane, demonstrating excellent fouling resistance. Furthermore, even if the contaminants in the raw filtrate block some of the surface pores as the filtration progresses, the large number of surface pores ensures a sufficient number of flow paths for the raw filtrate to pass through the porous membrane, making it easy to demonstrate excellent fouling resistance. 2 If this is the case, even if the filtrate is easily soiled, sufficient flow path can be secured until cleaning such as back pressure washing is performed. In addition, the filtrate reaches the pores on the surface of the porous membrane frequently and tends to flow straight through the porous membrane. The number of surface pores of the porous membrane is 290 / μm. 2 ~1500 pieces / μm 2 More preferably, the density is 350 particles / μm 2 ~1000 pieces / μm 2 It is particularly preferred that:

[0025] The number of surface pores of a porous membrane is the number of pores present within the surface when observing the surface of the porous membrane. To determine the number of surface pores of a porous membrane, the image obtained by observing the surface of the porous membrane with an SEM is binarized using the free software "ImageJ". When binarizing, create a background with 1 pixel using Subtract Background, then select the condition: RenyiEntropy in Threshold (binarization threshold). In the resulting binarized image, the number of pores in the observed range is determined using Analyze Particles. Number of surface pores per unit area [number / μm 2 To determine the area of ​​the hole, observe at least 1,000 holes and divide the number of holes by the total area of ​​the observed region.

[0026] Fig. 7 is a characteristic diagram showing the relationship between the number of surface pores and the dextran removal rate. The porous membranes of the prior art are shown in Fig. 7 as comparative examples (△ marks), but the greater the number of surface pores, the lower the dextran removal rate, and conversely, the higher the dextran removal rate, the fewer the number of surface pores, which is a trade-off. The porous membranes of the present invention are shown in Fig. 7 as examples (◯ marks), but the number of surface pores is 200 / µm. 2 As described above, it is clear that the dextran removal rate is defined as 60% or more. The porous membrane of the present invention overcomes the trade-off of conventional technologies and achieves both a large number of surface pores and a high dextran removal rate, thereby demonstrating excellent fouling resistance.

[0027] A porous membrane having an average surface pore diameter [nm] of 5.0 to 12.0 nm on the surface thereof prevents coarse contaminants and fine substances to be removed from the filtrate stock solution from penetrating into the porous membrane, and is therefore likely to exhibit high fouling resistance. A surface pore diameter [nm] of 12.0 nm or less prevents coarse substances to be removed from the filtrate stock solution (Stokes diameter: 13 nm or more) from penetrating into the porous membrane, and is therefore likely to exhibit high fouling resistance. A surface pore diameter [nm] of 5.0 nm or more allows fine substances (molecular weight: 10,000 Da or less) that are not to be removed but are preferably allowed to permeate sufficiently without being trapped in the porous membrane, and is therefore likely to exhibit high fouling resistance. A surface pore diameter [nm] of 5.0 to 9.0 nm is more preferred, and a surface pore diameter of 5.0 to 8.0 nm is particularly preferred.

[0028] The surface pore size is the diameter of the pores present in the plane when observing the surface of a porous membrane. To determine the surface pore size of a porous membrane, the image obtained by SEM observation of the surface of the porous membrane is binarized using the free software "ImageJ." When binarizing, create a background by setting 1 pixel in Subtract Background, and then select the condition: Renyi Entropy in Threshold (binarization threshold). In the resulting binarized image, select Area in Analyze Particles to determine the area of ​​each pore, and the diameter calculated by assuming each pore is a circle is used as the surface pore size. The average surface pore size is determined by averaging the diameters of more than 1,000 pores.

[0029] The value obtained by dividing the number of surface pores on the surface of the porous membrane (sometimes abbreviated as the number of surface pores) by the average surface pore diameter: X is 30 to 100 / μm 2 / nm, it is preferable because it prevents coarse contaminants and substances to be removed in the raw filtrate from penetrating into the porous membrane while ensuring a sufficient number of flow paths through which the raw filtrate passes through the porous membrane, thereby easily exhibiting excellent anti-fouling properties. A large X means that there are many small pores. Normally, when the pore size is small, the proportion of polymers constituting the porous membrane tends to increase so as to fill the pores, and therefore the number of pores tends to decrease. As a result, as shown in the comparative example (△) in Figure 8, as the surface pore size decreases, the number of surface pores also decreases, and there is a trade-off between the surface pore size and the number of surface pores.

[0030] After extensive research, the inventors discovered that small pores prevent coarse contaminants and substances to be removed in the raw filtrate from penetrating into the porous membrane, while large pores ensure a sufficient number of flow paths through which the raw filtrate passes through the porous membrane and allow for the dispersion of contaminants, making it easier to exhibit excellent fouling resistance. In other words, from the perspective of fouling resistance, it is preferable to achieve both a good (small) surface pore size and a good (large) number of surface pores. Since both the number of surface pores and the surface pore size have a positive correlation, which makes them good, and both contribute to fouling resistance, it is preferable to use the X value, which takes both into account, as an indicator of fouling resistance. Furthermore, since the number of surface pores is negatively correlated with the surface pore size, it is preferable to use the X value divided by the surface pore size as an indicator rather than using only the number of surface pores as an indicator.

[0031] Such a porous membrane with a large number of surface pores and a small surface pore diameter has not been obtained by conventional techniques. Figure 9 shows the relationship between the value X obtained by dividing the number of surface pores by the average surface pore diameter and the initial ratio of filtration flux. The closer the initial ratio of filtration flux F2 / F1 is to 1, the better the filtration performance over a long period of time. However, as shown in the examples and comparative examples, the porous membrane of the present invention has a value of X=30 to 100 pores / μm. 2 It has been shown that when the above ratio of F2 / F1 to filtration rate is met, good filtration with an F2 / F1 ratio of 0.50 or more can be maintained for a long period of time.

[0032] As will be described later, the porous membrane of the present invention has a relatively small self-diffusion coefficient of the polymer during the porous membrane formation process, i.e., it is difficult for the polymer to move, thereby suppressing the progression of excessive phase separation and coarsening, and the porous membrane is solidified in a state where there are many fine pores, making it easy to overcome the trade-off between pore size and pore number. X is 30 to 100 pores / μm, which is higher than the usual trade-off relationship. 2 / nm is preferable because it shows excellent stain resistance, and 32 to 80 particles / μm 2 / nm, more preferably 50 to 70 / μm 2 It is particularly preferred that the thickness is / nm.

[0033] A tortuosity ratio (R) of 1.0 to 8.0 allows fine contaminant components in the raw filtrate to pass through the porous membrane without clogging, and the porous membrane is likely to exhibit excellent fouling resistance, which is preferable. The tortuosity ratio, also known as the bending degree, refers to the degree to which the flow path through which the raw filtrate passes is curved. In other words, the lower the tortuosity ratio, the more linear the flow path, and the contaminant components pass through a shorter flow path, making them less likely to clog within the porous membrane. Conversely, as shown schematically in Figure 3, the higher the tortuosity ratio, the more curved the flow path is, and the contaminant components pass through a longer flow path, making them more likely to clog within the porous membrane. Note that Figure 3 is a schematic diagram showing the filtration status using a porous membrane of the prior art.

[0034] The tortuosity ratio (R) is generally calculated based on the Kozeny-Carman equation, and is described, for example, in Journal of the Japanese Association for Petroleum Technology (Vol. 75, No. 2 (March 2010) pp. 164-176). The tortuosity ratio (R) is preferably 1.0 to 6.0, and more preferably 1.0 to 5.5. Here, a tortuosity ratio (R) of 1.0 means that the pores are cylindrical and linear.

[0035] In the present invention, the tortuosity ratio: R is calculated using the following formula (1). R=(ε / 2k) 1 / 2 ·V / S ·······································Formula (1) ε: Porosity k: Permeability coefficient [m 2 ] V: Pore specific volume [m 3 / g] S: Specific surface area [m 2 / g] where k is the permeability coefficient at 25°C, and A is the membrane area at P [Pa]. 2 ] pure water permeability: Q [cm 3 / sec] using equation (2).

[0036] k=8.76×Q / A / P×10 -15 ...Equation (2) As mentioned above, the porosity (ε) can be determined by observing the cross section of the porous membrane from the surface to a thickness of 10 μm using an SEM and analyzing the binarized image. The specific surface area (S) and the specific pore volume (V) can generally be measured by the BET method using a commercially available nitrogen adsorption measurement device. In addition, the permeability coefficient (k[m 2 ] is 0.5 × 10 -17 m 2 ~5.0×10 -17 m 2 is preferable from the viewpoint of stably obtaining good properties of the porous film.

[0037] The porous membrane samples used for nitrogen adsorption measurements are prepared by cutting the surface of the porous membrane using a commercially available freezing microtome. The microtome moves the porous membrane a set distance, then cuts it by bringing the blade into contact with the membrane. The blade is set parallel to the surface of the porous membrane. First, the porous membrane is cut once by moving the blade closer to it at 0.5 μm intervals. Then, by moving the blade further to 10 μm and cutting once more, a surface section 10 to 10.5 μm thick can be extracted from the surface.

[0038] The permeability coefficient k is calculated based on formula (2) by filtering all the pure water through the porous membrane. The permeability coefficient k is 0.5 × 10 -17 ~5.0×10 -17 Preferably, it is 1.2 × 10 -17 ~5.0×10 -17 More preferably, it is 1.5×10 -17 ~5.0×10 -17 Here, the membrane area: A is the surface area of ​​the porous membrane that comes into contact with the raw filtrate. The pressure: P is the transmembrane pressure difference described above. As described above, the surface portion of the porous membrane, which is 10 μm thick from the surface, is the densest, so equation (2) is calculated assuming a surface thickness of 10 μm.

[0039] Within the surface of the porous membrane, the outermost surface from the surface to a thickness of 2 μm has an average cross-sectional pore diameter [nm] of 1 nm to 99 nm and the number of cross-sectional pores is 100 / μm.2 ~1000 pieces / μm 2 The porous membrane is preferred because it has a nanomesh structure, which tends to exhibit excellent anti-fouling properties. The cross-sectional pores refer to pores observed in the cross section of the outermost surface. The average cross-sectional pore diameter [nm] in the cross section of the outermost surface is as fine as 1 nm to 99 nm, so that coarse contaminant components and substances to be removed in the filtrate are easily captured at the outermost surface of the porous membrane and prevented from penetrating into the porous membrane. Furthermore, the number of cross-sectional pores is 100 / μm 2 ~1000 pieces / μm 2 Since there are many pores, the pores are formed in all directions, and the flow paths are easily connected in a straight line, meaning the tortuosity is low, and the porous membrane is less likely to become clogged when fine dirt components pass through it. Furthermore, even if some of the pores become clogged, a flow path for the raw filtrate to pass through the porous membrane can be secured, so the porous membrane is preferred because it tends to exhibit excellent fouling resistance.

[0040] Referring back to FIG. 1, the surface portion 102 of the porous membrane has a dense structure, with numerous pores that are connected vertically and horizontally and extend three-dimensionally. The pore diameter is less than 0.1 μm, and the number of pores forms an ultra-high density nanomesh structure. Fine dirt components 202 pass through any of the multiple interconnected pores without being trapped in the resin portion of the mesh structure. To allow fine components to pass through, the size of the pores in the porous membrane is adjusted to match the size of the fine components. However, by forming many pores densely, fine dirt components 202 are less likely to be trapped in the mesh. Therefore, the more pores there are, the better, which is preferable from the perspective of improved stain resistance and good permeability. Furthermore, since porous membranes have many pores, i.e., densely packed, the lower the tortuosity, the shorter the shortest distance formed by multiple pores. Therefore, even if some pores are blocked, a flow path can be secured by other pores nearby. This is preferable because the change in tortuosity corresponding to the shortest distance is small, making it easier to maintain permeability.

[0041] The nanomesh structure of the present invention is characterized by a finer pore size and a larger number of pores than a general mesh structure. A general mesh structure has a coarse pore size and a smaller number of pores. For example, the mesh structure shown in the figure of Patent Document 1 has a pore size of 1 to 2 μm and a pore count of 0.1 / μm. 2 The pore diameter of the mesh structure in the example of Patent Document 2 is as small as 300 nm, and the number of pores is several tens per μm 2 In contrast, in a nanomesh structure, the polymer resin portion forms a mesh, and as mentioned above, each pore in the mesh is very small, and there are many densely packed pores.

[0042] As described above, the general mesh structures known up to now have large pore diameters, which means that large contaminant components 201 penetrate deep into the porous membrane and easily clog, and the small number of pores makes it difficult for the flow paths to be connected in a straight line, resulting in a high degree of tortuosity and a long flow path through which fine contaminant components 202 pass, making them more likely to be trapped within the porous membrane and causing clogging. Figure 2 shows a schematic diagram of the filtration of a difficult-to-filter raw solution using such a porous membrane.

[0043] In Figure 2, coarse dirt components 201 are difficult to block at the surface layer 102, and tend to penetrate and remain in the porous membrane, resulting in clogging. Furthermore, fine dirt components 202 tend to be caught in the mesh structure with large pore diameters and high tortuosity, making it difficult for them to pass through the porous membrane and resulting in their retention. Compared to Figure 1, the coarse dirt components 201 and fine dirt components 202 foul the porous membrane, making it difficult for the membrane to maintain its filtering capacity over a long period of time.

[0044] The average cross-sectional pore diameter is more preferably 1.0 nm to 38 nm, and particularly preferably 1.0 to 36 nm. The number of cross-sectional pores is 120 to 800 / μm. 2 More preferably, the number of particles is 140 to 600 particles / μm 2 It is particularly preferred that:

[0045] The cross-sectional pore diameter of a porous membrane is calculated as follows. A cross-sectional sample for observation is prepared by embedding a porous membrane in a commercially available embedding medium for frozen tissue sectioning. Sections of the porous membrane with a thickness of 100 nm are then cut at low temperature using a freezing microtome and vacuum-dried at room temperature for 12 hours. The cross-section of the outermost surface of the porous membrane is observed using a transmission electron microscope (TEM). Images are then binarized using the free software "ImageJ." For binarization, select "Minimum" for the threshold. In the resulting binarized image, select "Area" under "Analyze Particles" to determine the area of ​​each pore. The diameter calculated assuming each pore is a circle is used as the cross-sectional pore diameter. The average cross-sectional pore diameter is calculated by averaging the diameters of 1,000 or more pores. The number of cross-sectional pores per unit area is calculated by dividing the number of cross-sectional pores by the total area of ​​the analyzed image. As with pore diameter, this is calculated by analyzing an image containing 1,000 or more pores.

[0046] The value obtained by dividing the number of cross-sectional pores on the outermost surface of the porous membrane by the average cross-sectional pore diameter: Y is 3 to 10 pores / μm 2 / nm prevents coarse contaminants and substances to be removed in the raw filtrate from penetrating into the porous membrane, while ensuring a sufficient number of flow paths through which the raw filtrate passes through the porous membrane, making it preferable for excellent fouling resistance to be exhibited. A large Y value means that there are many small pores. As explained above regarding surface pores, when the cross-sectional pore diameter is small, the number of cross-sectional pores tends to be small, and there is a trade-off between pore diameter and number of pores. After extensive research, the inventors have found that, even in the cross section of the outermost surface, small cross-sectional pores prevent coarse contaminants and substances to be removed in the raw filtrate from penetrating into the porous membrane, while many cross-sectional pores ensure a sufficient number of flow paths through which the raw filtrate passes through the porous membrane and can disperse the contaminants, making it easy to exhibit excellent fouling resistance. The number of cross-sectional pores and the cross-sectional pore diameter are correlated, and both contribute to fouling resistance, so it is preferable to use the Y value, which takes both into account, as an indicator of fouling resistance. In addition, since the number of cross-sectional pores is negatively correlated with the cross-sectional pore diameter, it is preferable to use the Y value obtained by dividing the number of cross-sectional pores by the cross-sectional pore diameter as an index rather than using only the number of cross-sectional pores as an index. The porous membrane of the present invention has a Y value of 3 to 10 pores / μm, which is higher than the usual trade-off relationship.2 / nm is preferable because it shows excellent stain resistance, and 4 to 10 particles / μm 2 / nm, more preferably 5 to 10 / μm 2 It is particularly preferred that the thickness is / nm.

[0047] At the outermost surface of the porous membrane, the standard deviation of the cross-sectional pore diameter [nm] is preferably 1.0 nm to 50 nm, which allows for uniform loading of fouling substances onto the porous membrane and tends to exhibit excellent fouling resistance. The standard deviation of the cross-sectional pore diameter is more preferably 1.0 nm to 35 nm. The standard deviation of the cross-sectional pore diameter can be calculated by binarizing and analyzing a TEM image of the cross section of the porous membrane as described above to obtain data on each cross-sectional pore diameter.

[0048] A standard deviation of the surface pore diameter [nm] of the surface portion of the porous membrane of 0.5 to 5.0 nm is preferable because it allows for uniform loading of fouling substances onto the porous membrane and tends to exhibit excellent fouling resistance. The smaller the standard deviation of the surface pore diameter, the smaller the pore diameter variation, but a standard deviation of 0.5 to 4.0 nm is more preferable, and a standard deviation of 0.5 to 3.0 nm is even more preferable. The standard deviation of the surface pore diameter can be calculated by binarizing and analyzing SEM images of the surface of the porous membrane as described above to obtain data on each surface pore diameter.

[0049] <About the manufacturing method of porous membrane> The porous membrane of the present invention comprises step (A): a step of dissolving a polymer in a solvent to obtain a polymer solution, and then step (B): a porous membrane formation step of solidifying the polymer solution in a non-solvent to form a porous membrane, and in the polymer solution obtained in step (A), the self-diffusion coefficient [m 2 / sec] is 0.8×10 -11 m 2 / sec~1.6×10 -11 m 2 / sec, the non-solvent used in step (B) contains 90 to 100% by weight of water, and the temperature of the non-solvent is 6°C to 45°C.

[0050] The type of polymer used in step (A) is not particularly limited, and specific examples include polysulfone resins, polyethersulfone resins, polyvinylidene fluoride resins, nylon, cellulose esters such as cellulose acetate and cellulose acetate propionate, fatty acid vinyl esters, polyvinyl alcohol, polyvinyl acetate, polyvinylpyrrolidone, polymers of acrylic acid esters or methacrylic acid esters such as ethylene oxide, propylene oxide, and polymethyl methacrylate, and copolymers thereof.

[0051] In particular, in order to use a porous membrane for long-term filtration, it is preferable to periodically chemically clean accumulated contaminants, and it is particularly preferable to include a polyvinylidene fluoride resin, which has excellent chemical resistance. Polyvinylidene fluoride resin refers to a vinylidene fluoride homopolymer or a vinylidene fluoride copolymer. Here, vinylidene fluoride copolymer refers to a polymer having a vinylidene fluoride residue structure. Polymers having a vinylidene fluoride residue structure are typically copolymers of vinylidene fluoride monomers and other fluorine-based monomers. Examples of such fluorine-based monomers include vinyl fluoride, tetrafluoroethylene, hexafluoropropylene, and trifluorochloroethylene. In the vinylidene fluoride copolymer, ethylene or other fluorine-based monomers other than the above may be copolymerized to the extent that the effects of the present invention are not impaired.

[0052] The polyvinylidene fluoride resin preferably accounts for 50% by weight or more, and particularly preferably 60% by weight or more, when the weight of the porous membrane is taken as 100%. The weight-average molecular weight of the polymer is preferably 50 to 1,000,000 Da, since this allows the self-diffusion coefficient (described later) to be relatively slow and easily controlled within an appropriate range. A mixture of multiple polymers may also be used. The solvent preferably contains a good solvent. Here, a "good solvent" refers to a solvent that can dissolve 5% by weight or more of the polymer even at low temperatures of 60°C or less. Examples of good solvents include N-methyl-2-pyrrolidone (hereinafter "NMP"), 2-pyrrolidone (hereinafter "2P"), ε-caprolactam (hereinafter "ε-CL"), dimethylacetamide, dimethylformamide, methyl ethyl ketone, acetone, tetrahydrofuran, tetramethylurea, trimethyl phosphate, or a mixture thereof. The good solvent preferably accounts for 40% by weight or more, and particularly preferably 60% by weight or more, of the solvent. By including a large amount of good solvent, the polymer chains spread in the polymer solution, and the self-diffusion coefficient (described later) is relatively slow, making it easy to control it within an appropriate range, which is preferable. Here, the "non-solvent" in step (B) refers to a solvent that does not dissolve or swell the polymer even when heated to a high temperature up to its boiling point. Examples of the non-solvent include water, hexane, pentane, benzene, toluene, methanol, ethanol, carbon tetrachloride, o-dichlorobenzene, trichloroethylene, ethylene glycol, diethylene glycol, triethylene glycol, propylene glycol, butylene glycol, pentanediol, hexanediol, and low-molecular-weight polyethylene glycol, and other aliphatic hydrocarbons, aromatic hydrocarbons, aliphatic polyhydric alcohols, aromatic polyhydric alcohols, chlorinated hydrocarbons, other chlorinated organic liquids, and mixtures thereof.

[0053] The polymer concentration (wt%) in the polymer solution is preferably equal to or higher than the entanglement concentration in order to control the self-diffusion coefficient within an appropriate range. More specifically, 10 to 40 wt% is preferred, 12 to 30 wt% is even more preferred, and 15 to 25 wt% is particularly preferred. We have found that a polymer concentration of 10 wt% or higher can relatively slow the self-diffusion coefficient and enable coagulation in a state where the pores in the porous membrane are minute and numerous. The polymer concentration is more preferably 12 wt% or higher, and particularly preferably 15 wt% or higher. Furthermore, by keeping the polymer concentration 40 wt% or lower, the proportion of the polymer in the porous membrane can be reduced, ensuring a sufficient number of pores. The polymer concentration is more preferably 30 wt% or lower, and particularly preferably 25 wt% or lower.

[0054] The porous membrane formation process in step (B), in which a polymer solution is solidified in a non-solvent to form a porous membrane, is a process of forming a porous membrane by so-called non-solvent-induced phase separation. When the polymer solution comes into contact with the non-solvent, the polymer cannot completely dissolve in the solvent, causing phase separation into a polymer-rich phase and a solvent-rich phase, and each phase coarsens as it coalesces with the surrounding identical phase.

[0055] Figure 4 is a schematic diagram showing the porous membrane formation process in step (B). Phase separation progresses in the polymer solution in the order of Figure 4(a) to (f), and the areas with high polymer concentration (polymer-rich phase 301) become coarse. During this phase separation and / or coarse-graining process, the exchange of solvent and non-solvent progresses, and when the non-solvent concentration exceeds a certain level, the polymer solidifies and the structure of the porous membrane is fixed. At this time, the solvent-rich phase 302 becomes the pores of the porous membrane.

[0056] As a result of intensive research, the inventors have found that in the production of the porous membrane of the present invention, the self-diffusion coefficient of the polymer is set to a relatively low range of 0.8 × 10 -11 m 2 / sec~1.6×10 -11 m 2 / sec, it was found that excessive phase separation and coarsening could be suppressed, and the polymer-rich phase 301 and the solvent-rich phase 302 could be solidified in a state where they existed in fine and large amounts. -11 m 2 / sec or more, the polymer and solvent have a diffusion coefficient sufficient for phase separation, making it easy to form pores. If the polymer diffusion is so slow that phase separation is difficult, pores are unlikely to form. -11 m 2 / sec or less, excessive phase separation and coarsening are suppressed, and pore coalescence is suppressed, making it easier to form many micropores. In other words, by appropriately controlling the self-diffusion coefficient of the polymer within a relatively low range, it is possible to form a porous film with many micropores. The self-diffusion coefficient is 0.8 × 10 -11 m 2 / sec~1.4×10 -11 m 2 / sec is more preferably 0.8×10 -11 m 2 / sec~1.1×10 -11 m 2 / sec is even more preferable.

[0057] One method for determining the self-diffusion coefficient is by all-atom molecular dynamics calculations. All-atom molecular dynamics calculations are a method for determining the trajectory of each atom by solving the equation of motion of a molecular ensemble system for all of the constituent atoms one by one. First, a polymer solution system is created to have the polymer concentration (weight %) that will actually be used. At this time, a single model polymer chain is modeled so that its molecular weight is 800 to 6000, and is 1 / 200 to 1 / 5 of the weight-average molecular weight of the polymer that will actually be used. Potential parameters used in molecular dynamics calculations can include known parameters such as those from DREIDING [SL Mayo, BD Olafson, WA Goddard III, J. Phys. Chem. 94, 8897 (1990)], GAFF [J. Wang, RM Wolf, JW Cladwell, PA Kollman, DACase, J. Comput. Chem. 25, 1157 (2004)], OPLS-AA [WL Jorgensen, DS Maxwell, Julian Tirado-Rives, J. Am. Chem. Soc. 118, 11225 (1996)], and CHARMM [BR Brooks, RE Bruccoleri, BD Olafson, DJ States, S. Swaminathan, M. Karplus, J. Comput. Chem. 4, 187 (1983)]. However, parameters that reproduce the density and cohesive energy, which are physical quantities that represent the aggregation state of a solution system, are preferred.

[0058] The inventors' investigation into the reproducibility of the aggregation state revealed that the use of GAFF or OPLS-AA is particularly preferable. A constant pressure-temperature ensemble is constructed by controlling the temperature at 25°C using the Nose-Hoover method [Hoover, W.G. Phys. Rev. A, 31, 1695 (1985)] and the pressure at 1 bar using the Andersen method [H.C. Andersen, J. Chem. Phys. 72, 2384 (1980)]. The short-range Lennard-Johns interactions are handled by applying a switch function starting from 1.0 nm and cutting off at 1.2 nm. Long-range electrostatic interactions are calculated using the Particle Mesh Ewald method. After performing molecular dynamics calculations in the constant pressure-temperature ensemble until the density becomes constant, the unit cell length is adjusted to achieve the average density, and an additional 11 ns of calculations are performed in the constant temperature ensemble. Using a 10 ns trajectory, the mean square displacement (MSD) of each atom in the polymer is determined, and the self-diffusion coefficient of the polymer is calculated using the following equation (4). At this time, the range of MSD and t used to calculate D must be confirmed so that the value obtained by dividing log(MSD) by log(t) is in the range of 0.9 to 1.1. When multiple types of polymers are mixed to make a polymer solution, the self-diffusion coefficient of the polymer in the polymer solution is determined by taking the weighted average of the self-diffusion coefficients of each polymer based on the weight percentage of the polymer. D = MSD / 6t Equation (4) D: self-diffusion coefficient t: time.

[0059] In the production of the porous membrane of the present invention, when the non-solvent used for coagulation contains 90 to 100 wt % water, coagulation is rapid and the self-diffusion coefficient of the polymer solution is likely to affect the rates of phase separation and coarsening. In other words, it is easy to achieve the effect of controlling the self-diffusion coefficient of the polymer solution to a relatively low range. Furthermore, when the temperature of the non-solvent is 6°C to 45°C, coagulation is rapid and the self-diffusion coefficient of the polymer solution is likely to affect the rates of phase separation and coarsening. The temperature of the non-solvent is more preferably 10°C to 35°C, and even more preferably 15°C to 30°C.

[0060] In the above-described step (A) of the production of the porous membrane of the present invention, the solvent preferably contains a hydrogen-bonding solvent having both hydrogen-bond donor and hydrogen-bond acceptor properties and a molecular weight of 500 Da or less, which facilitates controlling the self-diffusion coefficient of the polymer within an appropriate range. Having hydrogen-bond donor properties means having positively polarized hydrogen atoms, specifically, hydroxyl groups (OH groups), carboxyl groups (COOH groups), amino groups (NH groups), etc. Having hydrogen-bond acceptor properties means having lone electron pairs, specifically, carbonyl groups, alkoxy groups, cyano groups, etc. When the solvent has hydrogen-bond donor and hydrogen-bond acceptor properties, strong solvent-solvent hydrogen-bond interactions suppress the movement of the solvated polymer, making it easier to control the self-diffusion coefficient of the polymer within a relatively slow, appropriate range. The hydrogen bond donor and hydrogen bond acceptor solvents are not particularly limited, but specific examples include 2P, ε-CL, 1,3-dimethylurea, N-methylacetamide, hydantoin, 2-imidazolidinone, and DL-pyroglutamic acid.

[0061] In the above-mentioned step (A), the dissolved polymer preferably contains a polymer having hydrogen bond donor and / or hydrogen bond acceptor properties, which facilitates controlling the self-diffusion coefficient of the polymer within an appropriate range. When the polymer has hydrogen bond donor and / or hydrogen bond acceptor properties, it interacts with the solvent having hydrogen bond donor and hydrogen bond acceptor properties through hydrogen bonds, which facilitates controlling the self-diffusion coefficient of the polymer within a relatively slow, appropriate range. It is more preferable that the polymer having hydrogen bond donor and / or hydrogen bond acceptor properties account for 10% to 50% by weight of the porous membrane, from the viewpoint of controlling the self-diffusion coefficient of the polymer within an appropriate range.

[0062] Furthermore, when the molar number of hydrogen bond acceptor functional groups contained in the solvent in step (A) is divided by the molar number of hydrogen bond donor functional groups contained in the polymer having hydrogen bond donor and / or hydrogen bond acceptor properties, the hydrogen bonding between the polymer and the solvent is in an appropriate range, and the self-diffusion coefficient of the polymer is easily controlled to a relatively slow, appropriate range by setting the value to 1.0 to 12. This value is more preferably 2.0 to 9.0, and particularly preferably 5.4 to 8.0.

[0063] When the value obtained by dividing the number of moles of hydrogen bond donor functional groups contained in the solvent in step (A) by the number of moles of hydrogen bond acceptor functional groups contained in the polymer having hydrogen bond donor and / or hydrogen bond acceptor properties is 0.5 to 5.0, the hydrogen bonding between the polymer and the solvent is in an appropriate range, and the self-diffusion coefficient of the polymer is easily controlled to a relatively slow, appropriate range. This value is more preferably 1.5 to 2.3, and particularly preferably 1.8 to 2.3.

[0064] The porous membrane of the present invention may further comprise other layers. In this case, it is preferable that the porous membrane of the present invention is disposed on the surface. By disposing the porous membrane of the present invention on the surface, components contained in the filtrate are less likely to penetrate into the porous membrane, and high permeability can be maintained for a long period of time. The other layer is not particularly limited as long as it is a component that can be overlapped with the porous membrane to form a layer, but it is preferable that the other layer is a support. Here, the term "support" refers to a structure that physically reinforces the porous membrane and has a higher breaking strength than the porous membrane. To increase the breaking strength of the support, the breaking strength (breaking strength per unit area) of the support is preferably 3 MPa or more, more preferably 10 MPa or more. When the composite membrane in which the other layer and the porous membrane of the present invention are arranged in layers is in the form of a hollow fiber, the breaking strength of the support is preferably 300 gf or more, more preferably 800 gf or more. [Example]

[0065] The present invention will be described in detail below with reference to examples, but the present invention is not limited to these examples. First, measurement methods and evaluation methods are described below.

[0066] (i) Measurement of porosity The surface of the porous membrane was sampled using a commercially available freezing microtome for observation. The microtome allows the porous membrane to be moved a fixed distance before the blade comes into contact with the membrane and cutting it. The porous membrane, immersed in distilled water, was frozen at -20°C using a freezing microtome (Leica; Jung CM3000), and the blade was positioned perpendicular to the membrane's surface. The porous membrane was first cut once, moving closer to the blade at 30 μm intervals, to obtain sections. The sections were then vacuum-dried at 25°C for 12 hours. The cross-section of the porous membrane was observed using a SEM (Hitachi; SU-1510) at 10 μm intervals across the thickness of the porous membrane, and the porosity of each region was determined. To determine the porosity, the SEM images of the porous membrane cross-section were binarized using the free software "ImageJ." For binarization, we used Subtract Background to create a background of 1 pixel, then selected Persentile as the Threshold (binarization threshold). In the resulting binarized image, we selected Area in Analyze Particles to determine the area of ​​the pores, and divided this by the total area of ​​the observation region of the porous membrane to calculate the porosity as a percentage.

[0067] (ii) Permeability coefficient The entire amount of pure water at 25°C was filtered through the porous membrane, and the permeability coefficient was calculated based on equation (2): k [m 2 ] was calculated. Here, the membrane area: A [m 2 ] is the surface area of ​​the porous membrane in contact with the filtrate. The pressure: P [Pa] is the transmembrane pressure difference mentioned above. Here, the permeability coefficient is calculated for a thickness of 10 μm, taking into account the thickness [m] of the densest surface of the porous membrane. k=8.76×Q / A / P×10 -15 ...Equation (2).

[0068] (iii) Measurement of specific pore volume and specific surface area The surface of the porous membrane was sampled using a commercially available freezing microtome (Leica; Jung CM3000). The porous membrane was immersed in distilled water and frozen at -20°C using a freezing microtome (Leica; Jung CM3000), and the blade was placed parallel to the surface of the porous membrane. First, the porous membrane was cut once by moving the blade close to it at 0.5 μm intervals. Then, the moving distance was increased to 10 μm and another cut was made to sample a surface section 10 to 10.5 μm thick from the surface. The obtained surface section was analyzed by the BET method using a commercially available nitrogen adsorption measurement device (Microtrack Bell; BELSORP-mini II) to measure the specific pore volume: V [m 3 / g] and specific surface area S [m 2 / g] was measured.

[0069] (iv) Curvature ratio The porosity calculated in (i) for the region 10 μm thick from the surface of the porous membrane was used, and the permeability coefficient, specific pore volume, and specific surface area calculated in (ii) and (iii) were substituted into equation (1) to calculate the tortuosity: R. R=(ε / 2k) 1 / 2 ·V / S ···································Formula (1).

[0070] (v) Evaluation method for the removal rate of dextran with a weight-average molecular weight of 40,000 Da Dextran (manufactured by Aldrich; weight-average molecular weight 40,000 Da) was mixed with 1,000 ppm distilled water to prepare an aqueous dextran solution. The prepared aqueous dextran solution was supplied to a porous membrane at 25°C so that the transmembrane pressure difference was 10 kPa, and cross-flow filtration was performed at a cross-flow linear velocity of 1.0 m / sec, and the permeate was sampled. At the same time as the permeate was sampled, the aqueous dextran solution (stock solution) supplied to the porous membrane was also sampled. The refractive indexes of the permeate and stock solution were measured, and the removal rate: T (%) was calculated based on equation (3). T = {(Refractive index of the undiluted solution) - (Refractive index of the transmitted solution)} / (Refractive index of the undiluted solution) × 100 Equation (3).

[0071] (vi) Surface structure of the surface (pore size, number, standard deviation) The porous membranes were vacuum-dried at 25°C for 12 hours and then observed at 30,000–100,000 magnifications using a SEM (Hitachi High-Technologies Corporation; S-5500). Images of the porous membrane surface obtained using the SEM were binarized using the free software "ImageJ." For binarization, a 1-pixel subtraction was used to create a background, and the threshold (binarization threshold) condition: Renyi Entropy was selected. In the resulting binarized image, the area of ​​each surface pore was calculated by selecting "Area" in "Analyze Particles." The diameter calculated assuming each surface pore was a circle was used as the surface pore size. The average surface pore size was calculated by averaging the surface pore sizes of over 1,000 pores. Similarly, the standard deviation was calculated from each surface pore size data. The number of surface pores was divided by the total area of ​​the observed region to determine the number of surface pores per unit area.

[0072] (vii) Cross-sectional structure of the outermost surface (presence or absence of nanomesh structure, pore size, number, standard deviation) The porous membrane was embedded in a commercially available embedding medium for frozen tissue sectioning (OCT Compound, manufactured by Tissue-Tec). 100-nm-thick sections were then cut perpendicular to the surface of the porous membrane using a cryo-ultramicrotome (Leica FC7) at -40°C and vacuum-dried at room temperature for 12 hours. The cross-section of the outermost surface of the porous membrane was observed using a TEM (JEOL JEM-1400Plus), and images were binarized using the free software "ImageJ." The threshold was set to Minimum. The area of ​​each cross-sectional pore was calculated using the "Area" option in the "Analyze Particles" menu. The diameter of each cross-sectional pore was calculated assuming each pore was circular.

[0073] When calculating the average cross-sectional pore size, the cross-sectional pore sizes of 1,000 or more cross-sectional pores are averaged. The number of cross-sectional pores per unit area is calculated by dividing the number of cross-sectional pores by the total area of ​​the analyzed region. In this case, as with the surface pore size, the calculation is performed by analyzing an image containing 1,000 or more pores. Similarly, the standard deviation was calculated from the data on each cross-sectional pore size. Within the surface portion of the porous membrane, in the outermost surface portion up to a thickness of 2 μm from the surface, the average cross-sectional pore size [nm] in the cross section of the outermost surface portion is 1 nm to 99 nm, and the number of cross-sectional pores is 100 / μm. 2 ~1000 pieces / μm 2 When the mesh structure was larger than 100%, it was judged to have a nanomesh structure.

[0074] (viii) Self-diffusion coefficient of polymer The self-diffusion coefficient was determined by all-atom molecular dynamics calculations. A polymer solution system was prepared to match the actual polymer concentration (wt%). Each model polymer chain was modeled to have a molecular weight of 800–6000, 1 / 200–1 / 5 of the weight-average molecular weight of the actual polymer. GAFF2 was used as the potential parameter for the molecular dynamics calculations. A constant pressure-temperature ensemble was constructed by controlling the temperature at 25°C and the pressure at 1 bar. Short-range Lennard-Johns interactions were treated by applying a switch function starting from 1.0 nm and cutting off at 1.2 nm. Long-range electrostatic interactions were calculated using the Particle Mesh Ewald method. After performing molecular dynamics calculations in the constant pressure-temperature ensemble until the density was constant, the unit cell length was adjusted to achieve the average density, and an additional 11 ns of calculations were performed in the constant temperature ensemble. The mean square displacement (MSD) of each atom in the polymer was calculated using a 10-ns trajectory, and the self-diffusion coefficient was calculated using Equation (4) below. In this case, it was confirmed that the range of MSD and t used to calculate D was 0.9 to 1.1, where log(MSD) divided by log(t). When multiple types of polymers are mixed to make a polymer solution, the self-diffusion coefficient of each polymer was weighted and averaged based on the weight percentage of the polymer, and this was taken as the self-diffusion coefficient of the polymer in the polymer solution.

[0075] D = MSD / 6t Equation (4) D: self-diffusion coefficient t: time.

[0076] (ix) Filtration evaluation method for gelatin aqueous solution Gelatin (Gelatin Silver, manufactured by Nitta Gelatin Co., Ltd.) was mixed with 1000 ppm distilled water to prepare an aqueous gelatin solution. The prepared aqueous gelatin solution was supplied to a porous membrane at 60°C so that the transmembrane pressure difference was 120 kPa, and cross-flow filtration was performed at a cross-flow linear velocity of 1.0 m / sec, and the permeate was sampled. At the same time as the permeate was sampled, the aqueous gelatin solution (stock solution) supplied to the porous membrane was also sampled. The refractive indexes of the permeate and stock solution were measured, and the removal rate (%): T was calculated based on equation (5). ゼラチン The permeate volume was calculated as 20 L / m 2 Filtration flux (L / m 2 / h) was calculated. T ゼラチン = {(absorbance of the original solution at 292 nm) - (absorbance of the permeated solution at 292 nm)} / (absorbance of the original solution at 292 nm) × 100 ··· Equation (5).

[0077] (x) Apple juice filtration evaluation method Apple juice (manufactured by Ichiryu Co., Ltd.; fully ripe apple juice) was supplied to the porous membrane at 25°C so that the transmembrane pressure difference was 100 kPa, and the entire amount was filtered, and the permeate was sampled. At the same time as the permeate was sampled, the apple juice (raw solution) supplied to the porous membrane was also sampled. The turbidity of the permeate and the raw solution was measured, and the removal rate (%): T was calculated based on equation (6). りんご The permeate volume was calculated as 6 L / m 2 The filtration flux was calculated. T りんご = {(Turbidity of raw solution) - (Turbidity of permeated solution)} / (Turbidity of raw solution) × 100 Equation (6).

[0078] (xi) Filtration evaluation method for industrial wastewater Wastewater from a chemical factory (TOC: 30 mg / L, turbidity: 11 NTU) was supplied to the porous membrane at 25°C so that the transmembrane pressure difference was 100 kPa, and the total amount was filtered, and the permeate volume was measured.2 When this occurs, the reverse filtration permeation rate is 3 L / m so that the transmembrane pressure difference is 150 kPa. 2 Filtration and backfiltration were repeated until the filtration flux (F1) immediately after the start of filtration and the amount of permeate per unit area reached 400 L / m 2 The ratio of the filtration flux (F2) to the filtration flux immediately after the start of filtration after reaching the initial value (F2 / F1) was calculated and used as the initial filtration flux ratio. The closer the initial filtration flux ratio is to 1, the more the initial characteristics are maintained, and the porous membrane is less likely to clog even after long-term use. A guideline for maintaining good filtration is 0.50 or higher.

[0079] Example 1 The porous membrane used was a hollow fiber porous membrane containing a support obtained by the following manufacturing method. 38% by mass of PVDF (KF1300 manufactured by Kureha Corporation, weight-average molecular weight 350,000 Da) and 62% by mass of γ-butyrolactone were mixed and dissolved at 160°C to prepare a support membrane stock solution. This support membrane stock solution was discharged from a double-tube nozzle while accompanying an 85% by mass aqueous γ-butyrolactone solution as a hollow-portion-forming liquid. The discharged support membrane stock solution was solidified in a cooling bath containing an 85% by mass aqueous γ-butyrolactone solution at 20°C and placed 30 mm below the nozzle, to produce a hollow fiber support having a spherical structure.

[0080] A polymer solution having the composition shown in Table 1 was prepared by mixing 12% by mass of PVDF1 (Arkema's "Kynar" (registered trademark) 710, weight-average molecular weight 180,000 Da), 4.8% by mass of cellulose diacetate (Eastman's CA-398-3), 2.4% by mass of cellulose triacetate (Eastman's CA-436-80S), 68.8% by mass of NMP, and 12% by mass of 2P and stirring at 120°C for 4 hours.

[0081] Next, the polymer solution was uniformly applied to the outer surface of the hollow fiber support at 10 m / min (thickness: 50 μm). The support to which the polymer solution had been applied was taken up at 10 m / min, and 1 second after application, it was immersed in a coagulation bath of distilled water at 25°C for 10 seconds to coagulate, forming a porous membrane with a three-dimensional network structure. The evaluation results of the obtained porous membrane are shown in Table 1. The surface portion of the porous membrane from the surface to a thickness of 10 μm was denser than the interior. Figure 5 shows an image of the cross section of the outermost surface portion of the porous membrane observed with a TEM. From the cross-sectional pore diameter and the number of cross-sectional pores, it was determined that the porous membrane had a nanomesh structure. The tortuosity of the obtained porous membrane: R was 4.9, the dextran removal rate: T was 72%, and the number of surface pores: 444 / μm 2 The value obtained by dividing the number of surface pores by the average surface pore diameter: X is 62 / μm 2 / nm, which was good. Furthermore, (xi) the long-term stability was evaluated using the industrial wastewater filtration evaluation method, and the permeate volume was 400 L / m 2 The initial ratio (F2 / F1) of the filtration flux after (F2) to that immediately after the start of filtration (F1) was 0.63. The filtration flux was maintained even after long-term use. The evaluation results are shown in Table 1.

[0082] Example 2 A porous membrane was obtained in the same manner as in Example 1, except that 2P in the polymer solution was changed to ε-CL. The evaluation results of the porous membrane are shown in Table 1. The surface portion of the porous membrane from the surface to a thickness of 10 μm was denser than the interior. The tortuosity of the obtained porous membrane: R was 4.1, the dextran removal rate: T was 65%, and the number of surface pores: 291 / μm. 2 The value obtained by dividing the number of surface pores by the average surface pore diameter: X is 33 / μm 2 (xi) In the evaluation method for industrial wastewater filtration, the filtration flux ratio (F2 / F1) was 0.58, and the filtration flux was maintained even after long-term use.

[0083] Example 3 A porous membrane was obtained in the same manner as in Example 1, except that the composition ratio in the polymer solution was changed as shown in Table 1. The evaluation results of the porous membrane are shown in Table 1. The surface portion of the porous membrane from the surface to a thickness of 10 μm was denser than the interior. The tortuosity of the obtained porous membrane: R was 5.9, the dextran removal rate: T was 72%, and the number of surface pores: 329 / μm 2 The value obtained by dividing the number of surface pores by the average surface pore diameter: X is 46 / μm 2 (xi) In the evaluation method for industrial wastewater filtration, the filtration flux ratio (F2 / F1) was 0.55, and the filtration flux was maintained even after long-term use.

[0084] Example 4 A porous membrane was obtained by the same membrane production as in Example 1, except that the composition ratio in the polymer solution was changed as shown in Table 1. The porous membrane had a nanomesh structure. The evaluation results of the porous membrane are shown in Table 1. The surface portion of the porous membrane from the surface to a thickness of 10 μm was denser than the interior. The tortuosity of the obtained porous membrane: R was 4.9, the dextran removal rate: T was 69%, and the number of surface pores: 420 / μm 2 The value obtained by dividing the number of surface pores by the average surface pore diameter: X is 58 pores / μm 2 (xi) In the evaluation method for filtration of industrial wastewater, the filtration flux ratio (F2 / F1) was 0.62, and the filtration flux was maintained even after long-term use.

[0085] Example 5 (ix) The filtration performance of the porous membrane was evaluated using the filtration evaluation method for gelatin aqueous solution. The porous membrane used was that of Example 1. When the gelatin aqueous solution was filtered, the gelatin removal rate: T ゼラチン is 65%, and the filtration flux is 70 L / m 2 The results of the evaluation of the filtration of gelatin solution are shown in Table 3. The tortuosity of the porous membrane used in the evaluation was 4.9, the dextran removal rate was 72%, and the number of surface pores was 444 / μm. 2 The value obtained by dividing the number of surface pores by the average surface pore diameter: X is 62 / μm 2 / nm.

[0086] Example 6 (x) The filtering property of the porous membrane was evaluated using the apple juice filtering evaluation method. The porous membrane used was that of Example 1. When filtering apple juice, the removal rate of apple juice: T りんご is 99.9%, and the filtration flux is 37 L / m 2 The filtration rate of apple juice was excellent, at 1000 sachets / h. The results of the evaluation of apple juice filtration are shown in Table 4. The tortuosity of the porous membrane used in the evaluation was 4.9, the dextran removal rate was 72%, and the number of surface pores was 444 / μm. 2 The value obtained by dividing the number of surface pores by the average surface pore diameter: X is 62 / μm 2 / nm.

[0087] (Comparative Example 1) A porous membrane was obtained in the same manner as in Example 1, except that the solvent in the polymer solution was changed to NMP and the composition ratio was changed. The evaluation results of the porous membrane are shown in Table 2. The porous membrane did not have a nanomesh structure. The surface portion of the porous membrane, from the surface to a thickness of 10 μm, was denser than the interior, but the number of surface pores and the dextran removal rate (T) did not satisfy the requirements. (xi) In the evaluation method for filtration of industrial wastewater, the initial filtration flux ratio (F2 / F1) was 0.45, and after long-term use, the filtration flux decreased to less than half.

[0088] (Comparative Example 2) A porous membrane was produced in the same manner as in Comparative Example 1, except that the PVDF1 in the polymer solution was replaced with PVDF2 (Arkema: HSV900, melt viscosity 4700 Pa s) and the composition ratio was changed. The evaluation results of the porous membrane are shown in Table 2. The surface region of the porous membrane, from the surface to a depth of 10 μm, was denser than the interior, but the number of surface pores and the dextran removal rate (T) did not satisfy the requirements. (xi) In the evaluation method for industrial wastewater filtration, the initial filtration flux ratio (F2 / F1) was 0.40, and the filtration flux significantly decreased after prolonged use.

[0089] (Comparative Example 3) A porous membrane was obtained by the same procedure as in Comparative Example 1, except that PVDF3 (Solef9009 manufactured by Solvay) was used instead of PVDF1 in the polymer solution. The evaluation results of the porous membrane are shown in Table 2. The porous membrane did not have a nanomesh structure. The surface portion of the porous membrane, from the surface to a thickness of 10 μm, was denser than the interior, but the number of surface pores, the tortuosity (R), and the dextran removal rate (T) did not satisfy the requirements. (xi) In the method for evaluating the filtration of industrial wastewater, the initial filtration flux ratio (F2 / F1) was 0.40, and the filtration flux significantly decreased after long-term use.

[0090] Comparative Example 4 The porous membrane in Comparative Example 4 was an ST membrane manufactured by Synder. Figure 6 shows a TEM image of the cross section of the outermost surface of the porous membrane. The evaluation results are shown in Table 2. The porous membrane did not have a nanomesh structure. The surface portion of the porous membrane from the surface to a thickness of 10 μm was denser than the interior, but the number of surface pores, tortuosity (R), and dextran rejection rate (T) did not meet the requirements. (xi) In the industrial wastewater filtration evaluation method, the amount of permeated liquid was extremely small, making it difficult to measure the filtration flux, which was below the lower limit of measurement. It is believed that the extremely low filtration flux was due to the small number of surface pores and the excessively large dextran rejection rate (T).

[0091] (Comparative Example 5) A porous membrane was obtained in the same manner as in Comparative Example 3, except that the composition ratio in the polymer solution was changed. The composition ratio of 2P was 30 mass%. The evaluation results of the porous membrane are shown in Table 2. The surface portion of the porous membrane from the surface to a thickness of 10 μm was denser than the interior, but the number of surface pores, tortuosity (R), and dextran removal rate (T) did not satisfy the requirements. (xi) In the method for evaluating the filtration of industrial wastewater, the initial filtration flux ratio (F2 / F1) was 0.45, and the filtration flux significantly decreased after long-term use.

[0092] (Comparative Example 6) A porous membrane was obtained in the same manner as in Example 1, except that 2P in the polymer solution was changed to γ-butyrolactone. The evaluation results of the porous membrane are shown in Table 2. The surface region of the porous membrane, from the surface to a depth of 10 μm, was denser than the interior, but the average surface pore size, tortuosity (R), and dextran removal rate (T) did not satisfy the requirements. (xi) In the evaluation method for industrial wastewater filtration, the initial filtration flux ratio (F2 / F1) was 0.40, and the filtration flux significantly decreased after prolonged use.

[0093] (Comparative Example 7) (ix) The filtration performance of the porous membrane was evaluated using the filtration evaluation method for gelatin aqueous solution. The porous membrane used was the one used in Comparative Example 1. When the gelatin aqueous solution was filtered, the gelatin removal rate: T ゼラチン is 63%, and the filtration flux is 58 L / m 2 The number of surface pores on the porous membrane used for evaluation was 196 / μm. 2 The tortuosity ratio (R) was 6.4, and the dextran removal rate (T) was 55%.

[0094] (Comparative Example 8) (x) The filtering property of the porous membrane was evaluated using the apple juice filtering evaluation method. The porous membrane used was that of Comparative Example 4. When the apple juice was filtered, the removal rate of the apple juice: T りんご is 99.9%, and the filtration flux is 4 L / m 2 / h, the filtration flux was low and poor. The number of surface pores in the porous membrane used for evaluation was below the lower limit of observation, the tortuosity ratio (R) was 66.9, and the dextran removal rate (T) was 99.9%. The extremely low filtration flux is thought to be due to the small number of surface pores and the excessively high dextran removal rate (T).

[0095] (Comparative Example 9) A polymer solution was prepared by mixing 12% by mass of PVDF3 (Solef9009 manufactured by Solvay), 7% by mass of cellulose triacetate (CTA) (LT-35 manufactured by Daicel Corporation), and 81% by mass of NMP and stirring at 120°C for 4 hours.

[0096] Next, the polymer solution was uniformly applied to the outer surface of the hollow fiber support at 10 m / min (thickness: 50 μm). The support to which the polymer solution was applied was taken up at 10 m / min, and 1 second after application, it was immersed in a coagulation bath of distilled water at 15°C for 10 seconds to coagulate, forming a porous membrane with a three-dimensional network structure. The evaluation results of the obtained porous membrane are shown in Table 5. The surface portion of the porous membrane from the surface to a thickness of 10 μm was denser than the interior. The tortuosity of the obtained porous membrane: R was 6.0, the dextran removal rate: T was 71%, and the number of surface pores: 49 / μm 2 , the value obtained by dividing the number of surface pores by the average surface pore diameter: X is 6 / μm 2 / nm, and it was determined that it did not have a nanomesh structure based on the cross-sectional pore diameter and number of cross-sectional pores. (xi) In the filtration evaluation method for industrial wastewater, the initial filtration flux ratio (F2 / F1) was 0.42, and the filtration flux significantly decreased after long-term use.

[0097] (Comparative Example 10) A porous membrane was obtained in the same manner as in Comparative Example 1, except that the coagulation bath temperature was changed to 6°C. The evaluation results of the porous membrane are shown in Table 5. The surface portion of the porous membrane from the surface to a thickness of 10 μm was denser than the interior. The tortuosity of the obtained porous membrane: R was 13.2, the dextran removal rate: T was 70%, and the number of surface pores: 78 / μm 2 , the number of surface pores divided by the average surface pore diameter: X is 10 pores / μm 2 / nm, and (xi) in the filtration evaluation method for industrial wastewater, the initial filtration flux ratio (F2 / F1) was 0.37, and the filtration flux significantly decreased after long-term use.

[0098] (Comparative Example 11) A polymer solution was prepared by mixing 20% ​​by mass of polyethersulfone: PES (5900P manufactured by Sumika Excel Co., Ltd.), 4.1% by mass of polyvinylpyrrolidone: PVP (K-30 manufactured by Nippon Shokubai Co., Ltd.), 4.1% by mass of polyethylene glycol: PEG (PEG300 manufactured by Wako Pure Chemical Industries, Ltd.), 35.9% by mass of 2P, and 35.9% by mass of NMP, and stirring the mixture at 120°C for 4 hours.

[0099] Next, the polymer solution was uniformly applied to the outer surface of the hollow fiber support at 10 m / min (thickness: 50 μm). The support to which the polymer solution was applied was taken up at 10 m / min, and 1 second after application, it was immersed in a coagulation bath of distilled water at 25°C for 30 seconds to coagulate, forming a porous membrane with a three-dimensional network structure. The evaluation results of the obtained porous membrane are shown in Table 5. The surface portion of the porous membrane from the surface to a thickness of 10 μm was denser than the interior. The dextran removal rate of the obtained porous membrane: T was 24%, and the number of surface pores: 2 / μm 2 The value obtained by dividing the number of surface pores by the average surface pore diameter: X is 0.01 pores / μm 2 / nm, and (xi) in the industrial wastewater filtration evaluation method, the amount of permeate was extremely small, making it difficult to measure the filtration flux, which was below the lower limit of measurement. The extremely low filtration flux is thought to be due to the small number of surface pores and the low dextran removal rate: T.

[0100] (Comparative Example 12) A polymer solution was prepared by mixing 10% by mass of cellulose diacetate (CDA) (Daicel Corporation: L-70), 4.1% by mass of polyvinylpyrrolidone (PVP) (Nippon Shokubai Co., Ltd.: K-30), 4.1% by mass of polyethylene glycol (PEG) (Wako Pure Chemical Industries, Ltd.: PEG300), 40.9% by mass of 2P, and 40.9% by mass of NMP at 120°C for 4 hours.

[0101] Next, the polymer solution was uniformly applied to the outer surface of the hollow fiber support at 10 m / min (thickness: 50 μm). The support to which the polymer solution was applied was taken up at 10 m / min, and 1 second after application, it was immersed in a coagulation bath of distilled water at 25°C for 30 seconds to coagulate, forming a porous membrane with a three-dimensional network structure. The evaluation results of the obtained porous membrane are shown in Table 5. The surface portion of the porous membrane from the surface to a thickness of 10 μm was denser than the interior. The dextran removal rate of the obtained porous membrane: T was 43%, and the number of surface pores: 9 / μm 2 The value obtained by dividing the number of surface pores by the average surface pore diameter: X is 0.4 / μm 2 / nm, and (xi) in the filtration evaluation method for industrial wastewater, the initial filtration flux ratio (F2 / F1) was 0.47, and the filtration flux significantly decreased after long-term use.

[0102] 10 shows the relationship between the self-diffusion coefficient and the initial ratio of filtration flux (F2 / F1) for the example and the comparative example. -11 m 2 / sec~1.6×10 -11 m 2 / sec, the initial ratio of filtration flux was 0.50 or more, and good filtration performance was maintained even after long-term operation.

[0103] [Table 1]

[0104] [Table 2]

[0105] [Table 3]

[0106] [Table 4]

[0107] [Table 5] [Explanation of symbols]

[0108] 101 Porous membrane 102 Surface part 103 Internal 201 Coarse dirt components 202 Fine dirt components 300 Polymer Solution 301 Polymer-rich phase 302 Solvent-rich phase FL Filtration direction

Claims

1. On at least one surface, the surface portion from the surface to a thickness of 10 μm is denser than the interior, the dextran removal rate (T) of which has a weight-average molecular weight of 40,000 Da is 60% to 95%, and the number of pores observed on the surface of the surface portion (hereinafter referred to as surface pores) per unit area is 200 / μm ２ ~2000 pieces / μm ２ The average value of the surface pore diameter [nm] of the surface portion is 5.0 nm to 12 nm, and the number of pores observed in a cross section of the outermost surface portion (hereinafter referred to as cross-sectional pores) per unit area is 100 / μm in the outermost surface portion up to a thickness of 2 μm from the surface of the surface portion. ２ ~1000 pieces / μm ２ and having a nanomesh structure in which the average cross-sectional pore diameter [nm] is 1 nm to 99 nm.

2. The number of surface pores per unit area [number / μm ２ ] by the average surface pore diameter [nm] of the surface portion: X is 30 to 100 / μm ２ The porous membrane according to claim 1, wherein the pore size is 1 / nm.

3. The number of cross-sectional holes per unit area [number / μm ２ ] by the average cross-sectional pore diameter [nm] of the outermost surface portion: Y is 3 to 10 pores / μm ２ The porous membrane according to claim 1 or 2, wherein the pore size is 1 / nm.

4. 3. The porous membrane according to claim 1, wherein the standard deviation of the cross-sectional pore diameter [nm] of the outermost surface portion is 1.0 nm to 50 nm.

5. The porous membrane according to claim 1 or 2, wherein the standard deviation of the surface pore diameter [nm] of the surface portion is 0.5 nm to 5.0 nm.

6. A method for filtering a liquid using the porous membrane according to claim 1 or 2.

Images