Carbon Membranes for Fluid Separation

Abstract

An object of the present invention is to suppress defects in a carbon membrane for fluid separation in which a dense carbon layer is formed on a porous carbon support. The present invention provides a carbon membrane for fluid separation in which a dense carbon layer is formed on a porous carbon support, wherein the ratio of the content of silicon atoms to the total content of carbon atoms and silicon atoms at the center position in the membrane thickness direction of the porous carbon support is X (atomic%), and the ratio of the content of silicon atoms to the total content of carbon atoms and silicon atoms at a position 3 μm from the interface with the dense carbon layer of the porous carbon support toward the porous carbon support side is Y (atomic%). When X < Y, it is a carbon membrane for fluid separation being provided.

Description

[Technical field]

[0001] The present invention relates to carbon membranes for fluid separation. [Background technology]

[0002] Membrane separation is used as a means of selectively separating and purifying specific components from various mixed gases and liquids. Membrane separation is attracting attention because it is more energy-efficient than other fluid separation methods. Types of separation membranes that have been proposed include organic polymer membranes such as polyimide membranes and cellulose acetate membranes, and inorganic membranes such as zeolite membranes, silica membranes, and carbon membranes.

[0003] Among them, carbon membranes have a molecular sieve effect that can separate molecules according to their size and have excellent heat resistance and durability, and therefore various separation membranes having a separation layer made of a dense carbon layer have been proposed. For example, a porous ceramic composite hollow fiber membrane (see Patent Document 1, for example) formed by supporting a thin carbon film on the surface of a porous ceramic hollow fiber membrane, a hollow fiber carbon membrane (see Patent Document 2, for example) that includes a hollow fiber-shaped first carbon membrane and a second carbon membrane provided on the outer surface of the first carbon membrane, the second carbon membrane containing a metal element and a sulfur element, and a carbon membrane for fluid separation (see Patent Document 3, for example) that includes a core layer having a cocontinuous porous structure and a skin layer formed around the core layer that does not substantially have a cocontinuous porous structure have been proposed. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 08-299769 [Patent Document 2] JP 2013-63415 A [Patent Document 3] International Publication No. 2016 / 13676 Summary of the Invention [Problem to be solved by the invention]

[0005] Although the carbon membrane can improve separation and permeation performance, in recent years, separation and purification at high pressures are required. For example, in natural gas purification plants, it is necessary to separate and remove carbon dioxide, which is an impurity, from methane gas, which is the main component. Since the larger the pressure difference between the upstream and downstream sides of the separation membrane, the higher the permeation rate, separation and purification at high gas pressures of several MPa or more is required from the viewpoint of energy efficiency. In addition, in the chemical industry, membrane separation methods have begun to be used in processes for separating and purifying water, which is an impurity, from alcohol and acetic acid, and separation and purification at high pressures are required to improve the permeation flow rate of the substance to be separated.

[0006] However, the separation membranes having a separation layer made of a conventional carbon layer described in Patent Documents 1 to 3 are prone to defects such as pinholes and cracks due to the effects of stress caused by expansion and contraction during the production process and the separation and purification process, defects derived from the hollow fiber membrane (core layer), and defects due to insufficient interfacial affinity between the hollow fiber membrane (core layer) and the separation membrane, and have a problem in that these defects are more likely to occur under high-pressure conditions.

[0007] In particular, compared with supports made of ceramics such as alumina, supports made of carbon are more susceptible to expansion and contraction during the manufacturing process, and the resulting stress is more likely to cause defects in the support or carbon film. In addition, in the manufacturing process of carbon supports, the supports are fused together due to contact with each other, and defects occur when they are peeled off, and these defects cause defects in the carbon film, which is an issue.

[0008] If defects that occur in a separation membrane made of a carbon layer are larger than the molecules of the fluid to be separated, the fluid to be separated will leak through the defects, so it is necessary to suppress the defects. [Means for solving the problem]

[0009] The present invention for solving the above problems is a carbon membrane for fluid separation in which a dense carbon layer is formed on a porous carbon support, where the ratio of the content of silicon atoms to the total content of carbon atoms and silicon atoms at the central position in the membrane thickness direction of the porous carbon support is X (atomic %), and the ratio of the content of silicon atoms to the total content of carbon atoms and silicon atoms at a position 3 μm on the porous carbon support side from the interface with the dense carbon layer of the porous carbon support is Y (atomic %). When Y is defined as such, it is a carbon membrane for fluid separation where X < Y.

Effect of the Invention

[0010] According to the present invention, in a carbon membrane for fluid separation in which a dense carbon layer is formed on a porous carbon support, defects can be suppressed.

Brief Description of the Drawings

[0011] [Figure 1] It is an example of a schematic cross-sectional view of a carbon membrane for fluid separation. [Diagram 2] It is an example of an enlarged photograph of a co-continuous porous structure. [Diagram 3] It is a diagram of a method for determining the measurement location of X. [Figure 4] It is a diagram of a method for determining the measurement locations of Y and Z. [Diagram 5] It is a diagram of a method for determining the central position in the membrane thickness direction of the dense carbon layer.

Embodiments for Carrying Out the Invention

[0012] Hereinafter, in this specification, "~" indicates a range including the numerical values at both ends thereof.

[0013] <Carbon Membrane for Fluid Separation> FIG. 1 shows an example of a schematic cross-sectional view of the carbon membrane for fluid separation of the present invention. The carbon membrane for fluid separation of the present invention has a structure in which a dense carbon layer 2 is formed on a porous carbon support 1. In FIG. 1, the carbon membrane for fluid separation in which the porous carbon support 1 and the dense carbon layer 2 are formed on its outer surface is in the form of a hollow fiber, but its shape is not limited to this, and it may be in the form of a solid fiber or a film.

[0014] [Porous Carbon Support] A porous carbon support (hereinafter, the porous carbon support may be simply referred to as a "support") is a base material for maintaining the shape of a dense carbon layer having a function as a separation membrane. Since the support is formed of a carbon material, it has high heat resistance and chemical resistance compared to organic polymer supports. Here, the carbon support refers to a support having the maximum ratio of carbon atoms to all atoms contained therein. Specifically, the ratio of the carbon atom content to the total content of all atoms at the center position in the film thickness direction of the support is the maximum in the content ratio of each atom. From the viewpoint of improving the heat resistance and chemical resistance of the fluid separation membrane, the ratio of the carbon atom content to the total content of all atoms at the center position in the film thickness direction of the support is preferably 60 atomic % or more, more preferably 65 atomic % or more. On the other hand, from the viewpoint of increasing flexibility and improving handling, the ratio of the carbon atom content is preferably 95 atomic % or less, more preferably 85 atomic % or less.

[0015] The carbon atom content relative to the total content of all atoms at the center position in the thickness direction of the support can be determined by performing elemental analysis using an energy dispersive X-ray analyzer (EDX) at an acceleration voltage of 15 keV. 10 points are randomly measured from the center position in the thickness direction of the support determined by the method described below, and the average value of the ratio of the carbon atom content relative to the total content of all atoms obtained from the obtained values ​​is taken as the ratio of the carbon atom content relative to the total content of all atoms at the center position in the thickness direction of the support. However, if the support is solid, the center of gravity of the support cross section is measured as the center position in the thickness direction of the support, and if no other points to be scanned are obtained within one cross section, another cross section is prepared and the above operation is repeated, and the average value obtained from the measured values ​​of 10 points is used to calculate the ratio of the carbon atom content.

[0016] The center position in the film thickness direction of the support refers to the midpoint (point C in FIG. 3) of a line segment connecting two points (point A in FIG. 3) randomly selected from the interface 7 between the dense carbon layer 5 and the support 6, which is determined by exposing the cross section of the support by a cross-section polisher method (CP method) and using a method described later, and a point (point B in FIG. 3) on the other support surface other than the support surface that forms the interface with the dense carbon layer, which is the shortest distance from the point on the interface. However, when the support is solid, the center position of the gravity of the support cross section is taken as the center position of the film thickness.

[0017] The interface between the support and the dense carbon layer is defined as the interface between a layer where clear pores are observed and a layer where they are not observed. Specifically, for a cross section formed by the CP method, a scanning electron microscope (SEM) is used to move the field of view from the dense carbon layer to the support at a magnification of 1±0.1 (nm / pixel), and the edge of the first pore observed on the dense carbon layer side is defined as the interface between the dense carbon layer and the support.

[0018] In addition, since the support has a porous structure, it also serves as a flow path for fluids such as gases and liquids. Examples of the porous structure include various porous structures such as a closed cell structure and a continuous porous structure. If it is a closed cell structure, the cross-sectional compressive strength is improved. On the other hand, if it is a continuous porous structure, the pressure loss during the permeation of the fluid is reduced, and the permeation speed of the fluid is improved, which is preferable.

[0019] The porous structure of the support is preferably a co-continuous porous structure, which is one form of a continuous porous structure. The co-continuous porous structure is a structure in which the branches (carbon parts) and the pores (void parts) are each continuous and regularly intertwined in a three-dimensional manner. FIG. 2 shows an example of an enlarged photograph of the co-continuous porous structure. As shown in FIG. 2, the structure has a branch part 3 and a void part 4 that are each continuous in the depth direction. With such a structure, the branches each support the structure, dispersing stress throughout the support, improving the cross-sectional compressive strength and suppressing damage even when the fluid supplied is at high pressure.

[0020] The porous structure of the support can be observed by magnifying a cross section of a support sufficiently cooled in liquid nitrogen and cut with tweezers using a scanning electron microscope. The cocontinuous porous structure is distinct from irregular structures such as a structure in which individual particles are aggregated and connected, or a structure formed by voids generated by removing aggregated and connected template particles and the surrounding skeleton.

[0021] The porous structure of the support is preferably a highly uniform structure having a periodic structure, and the structural period is preferably 0.002 μm or more and 10 μm or less. If the structural period is 0.002 μm or more, the pressure loss when a fluid passes through the voids is reduced, and the permeation rate of the fluid is improved. The structural period is more preferably 0.01 μm or more, and even more preferably 0.05 μm or more. On the other hand, if the structural period is 10 μm or less, the cross-sectional compressive strength is improved. The structural period is more preferably 8 μm or less.

[0022] The structural period of the porous structure is calculated by the following formula using the value of the scattering angle 2θ at the position of the peak top of the scattering intensity obtained by incident X-rays on the carbon membrane for fluid separation of the present invention and scattering them at a small angle.

[0023]

number

[0024] P: Structural period (μm), λ: Wavelength of incident X-ray (μm) If the structural period of the support is too large to observe scattering at small angles, the structural period is obtained by X-ray computed tomography (X-ray CT). Specifically, a three-dimensional image taken by X-ray CT is Fourier transformed, and the circular average of the two-dimensional spectrum is taken to obtain a one-dimensional spectrum. The characteristic wavelength corresponding to the position of the peak top in the one-dimensional spectrum is determined, and the structural period of the porous carbon support is calculated from its reciprocal.

[0025] In addition, in the analysis of the above structural period, the structural period of the dense carbon layer is outside the above range and does not affect the analysis. Therefore, the structural period of the porous carbon support is the same as the value of the structural period calculated by measuring the entire carbon membrane for fluid separation.

[0026] Furthermore, the high uniformity of the porous structure reduces the occurrence of starting points for fracture of the carbon membrane for fluid separation, and has the effect of dispersing stress due to compression, etc., throughout the carbon membrane, thereby improving the cross-sectional compressive strength. In addition, the improved flexibility allows the bending radius to be reduced.

[0027] The uniformity of the porous structure can be evaluated by the half-width of the scattering intensity peak when X-rays are incident on the carbon membrane for fluid separation of the present invention. Specifically, in a graph in which the horizontal axis represents the scattering angle 2θ and the vertical axis represents the scattering intensity, the smaller the half-width of the scattering intensity peak, the higher the uniformity. The half-width of the peak is preferably 5° or less, more preferably 1° or less, and even more preferably 0.1° or less.

[0028] The half-width of the peak in the above description refers to the width of the peak at the midpoint (point C) of the line segment connecting points A and B, where the apex of the peak is point A, a straight line parallel to the vertical axis of the graph is drawn from point A, and the intersection point of the line with the baseline of the spectrum is point B. Note that the width of the peak here refers to the width on a line that is parallel to the baseline and passes through point C.

[0029] The average porosity of the porous structure of the support is preferably 10% or more and 80% or less. The average porosity is determined by observing the cross section of the support, which is precisely formed by the CP method with the embedded sample, at a magnification of 1±0.1 (nm / pixel) and a resolution of 700,000 pixels or more, and setting a region of interest from the image to be 512 pixels square. Then, the cross-sectional area is calculated as A, the total area of ​​the pores is B, and the value is calculated from the arithmetic average value of 20 randomly selected cross sections. Note that when the carbon membrane for fluid separation is a hollow fiber, the hollow portion is not included in the average porosity.

[0030] Average porosity (%)=B / A×100 The higher the average porosity, the smaller the pressure loss as a fluid flow path, and the higher the fluid permeation rate. Therefore, the average porosity is more preferably 15% or more, and even more preferably 18% or more. On the other hand, the lower the average porosity, the higher the cross-sectional compressive strength, and the higher the fluid permeation rate under high pressure conditions. Therefore, 75% or less is more preferable, and 70% or less is even more preferable.

[0031] In addition, if the average pore diameter of the support is large, the pressure loss decreases and the fluid permeation rate increases. Therefore, the average pore diameter is preferably 30 nm or more, more preferably 50 nm or more. On the other hand, if the average pore diameter is small, the cross-sectional compressive strength increases. Therefore, the average pore diameter is preferably 3,000 nm or less, more preferably 2,500 nm or less.

[0032] The average pore diameter of the support is the value obtained by analyzing the carbon membrane for fluid separation by mercury porosimetry. In the mercury porosimetry, the pore volume and specific surface area are calculated from the pressure when mercury is infiltrated into the pores by applying pressure and the amount of mercury infiltrated. Next, the pore radius or diameter is calculated assuming that the pores are cylindrical from the relationship between the pore volume and the specific surface area. In the mercury porosimetry, a pore diameter distribution curve of 5 nm to several hundred μm can be obtained, and the pore diameter at the peak top is taken as the average pore diameter of the support. Note that the dense carbon layer does not have pores of 5 nm or more, so the average pore diameter of the carbon membrane for fluid separation is substantially the same as the average diameter of the pores of the support. Note that even when the support is in the form of a solid thread, a porous structure is formed in the support.

[0033] In the present invention, the shape of the support is preferably fibrous or film-like. Fibrous refers to a support having a ratio of fiber length L to average fiber diameter D (aspect ratio L / D) of 100 or more, and examples thereof include hollow fibers having voids (hollow portions) of approximately the same diameter continuously formed in the fiber axial direction (longitudinal direction), and solid fibers without hollow portions. If the support is fibrous, the membrane area per unit volume of the separation module can be made larger than that of a film, making the separation module more compact.

[0034] When the support is fibrous, its average diameter can be set arbitrarily. If the average diameter is large, pressure loss between the upstream side and downstream side (permeation side) of the membrane is unlikely to occur, and it is easy to ensure the differential pressure required for fluid permeation. Therefore, the average diameter is preferably 100 μm or more. On the other hand, if the average diameter is small, bending rigidity is improved and the membrane area per unit volume in the separation module becomes large. Therefore, the average diameter is preferably 500 μm or less. In addition, the cross-sectional shape of the fiber is arbitrary, and examples thereof include a round cross section, a multi-leaf cross section such as a triangular cross section, a flat cross section, and a hollow cross section. A round cross section is preferable because it has a high cross-sectional compressive strength.

[0035] When the support is in the form of a hollow fiber, the area ratio of the cross-sectional area C of the hollow portion to the cross-sectional area A of the support (hollow area ratio: C / A) is preferably 0.001 or more and 0.7 or less. The larger the hollow area ratio, the lower the pressure loss and the higher the fluid permeation rate. Therefore, the hollow area ratio is more preferably 0.01 or more, and even more preferably 0.05 or more. On the other hand, the smaller the hollow area ratio, the higher the cross-sectional compressive strength. Therefore, the hollow area ratio is more preferably 0.6 or less.

[0036] When the hollow area ratio is within the above range, the cross-sectional compressive strength and the fluid permeation rate are well balanced. The cross-sectional area A of the support includes the cross-sectional area C of the hollow portion. In order to achieve both the cross-sectional compressive strength and the fluid permeation rate, the support may have a plurality of hollow portions, and in that case, the sum of the cross-sectional areas of the hollow portions is defined as the cross-sectional area C of the hollow portions.

[0037] Furthermore, in the case of a hollow fiber shape, the presence of hollow spaces during carbonization to form a dense carbon layer absorbs stress caused by carbonization shrinkage of the carbonizable resin, making it less likely to develop defects such as cracks, which is preferable.

[0038] The thicker the membrane of the hollow fiber support, the better the cross-sectional compressive strength and handling. Therefore, the membrane thickness of the support is preferably 10 μm or more, more preferably 30 μm or more. On the other hand, the thinner the membrane of the support, the better the permeability and flexibility, so the membrane thickness of the support is preferably 200 μm or less, more preferably 100 μm or less.

[0039] On the other hand, when the support is in the form of a film, the thickness of the support is not limited. If the support is thick, the handling property is improved. Therefore, the thickness is preferably 10 μm or more. Also, if the support is thin, the bending rigidity is improved and the support is less likely to break. Therefore, the thickness is preferably 5,000 μm or less.

[0040] The film thickness of the support is defined as the average value of the lengths of the line segments connecting one point randomly selected from the interface between the dense carbon layer and the support determined by the above-mentioned method and a point on the other surface of the support other than the surface that forms the interface with the dense carbon layer that is the shortest distance from the point on the interface, measured at 10 randomly selected points.

[0041] Through investigations by the present inventors, it has been found that in a carbon membrane for fluid separation, by including silicon in the vicinity of the interface between the support and the dense carbon layer, defects are less likely to occur in the dense carbon layer. Although the reason for this is not clear, it is believed that the inclusion of silicon in the vicinity of the interface between the support and the dense carbon layer improves the surface condition of the interface between the support and the dense carbon layer, thereby improving the wettability of the support during the formation of the dense carbon layer, suppressing the generation of coarse pores, and suppressing shrinkage and expansion during the manufacturing process, as well as suppressing defects in the support due to fusion and peeling between supports.

[0042] Therefore, in the present invention, when the ratio of the content of silicon atoms to the total content of carbon atoms and silicon atoms at the center position in the film thickness direction of the porous carbon support is X (atomic %), and the ratio of the content of silicon atoms to the total content of carbon atoms and silicon atoms at a position 3 μm from the interface with the dense carbon layer of the porous carbon support on the porous carbon support side is Y (atomic %), it is characterized in that X < Y. In the present invention, as an index of the content of silicon atoms unevenly distributed near the interface with the dense carbon layer of the porous carbon support, attention is paid to the content of silicon element at a position 3 μm from the interface with the dense carbon layer of the porous carbon support on the porous carbon support side, and the relationship with the content of silicon atoms at the center position in the film thickness direction, which is an index of the content of silicon atoms in the porous carbon support other than the interface with the dense carbon layer, is specified. When the relationship between X and Y described above satisfies X < Y, it means that the ratio of the content of silicon atoms in the vicinity of the interface with the dense carbon layer is higher than that at the center of the porous carbon support, that is, silicon atoms are unevenly distributed near the interface with the dense carbon layer. From the viewpoint of further suppressing defects, the difference (Y - X) between Y and X is preferably 0.01 or more, more preferably 0.10 or more, and even more preferably 1 or more. On the other hand, from the viewpoint of suppressing the peeling between the support and the dense carbon layer and further suppressing defects, Y - X is preferably 15 or less, more preferably 10 or less, and even more preferably 8 or less.

[0043] From the viewpoint of improving chemical resistance, compression resistance, and permeability, X is preferably 10 or less, and more preferably 5 or less. Also, from the viewpoint of further suppressing defects, Y is preferably 0.01 or more, more preferably 1 or more, and even more preferably 5 or more. On the other hand, from the viewpoint of increasing the adhesion between the support and the dense carbon layer, suppressing the peeling between the support and the dense carbon layer, and further suppressing defects, Y is preferably 40 or less, more preferably 20 or less, and even more preferably 10 or less.

[0044] Means for making X < Y, that is, means for unevenly distributing silicon near the interface, include, for example, in the production method of the present invention described later, a method of applying a silicon-containing component by a preferable method described later, and the like.

[0045] It is preferable that metal atoms are contained in the vicinity of the interface between the support and the dense carbon layer together with silicon atoms, so that defects can be further suppressed. As the metal atoms, alkali metals and alkaline earth metals are preferable. Two or more of these may be contained. Among these, sodium and calcium are more preferable.

[0046] The ratio Z (atomic %) of the metal atom content to the total content of carbon atoms and metal atoms at a position 3 μm away from the interface between the porous carbon support and the dense carbon layer is used as an index of the metal atom content near the interface, and Z is preferably greater than 0, and more preferably 0.0010 or greater.

[0047] An example of a method for adjusting Z to be within the above range is a method in which a metal-containing component is provided by a preferred method described below in the production method of the present invention described below.

[0048] The carbon atom and silicon atom contents at the center position in the film thickness direction of the porous carbon support can be determined by performing elemental analysis using an energy dispersive X-ray analyzer (EDX) at an acceleration voltage of 15 keV. The carbon atom and silicon atom contents at the center position in the film thickness direction of the support can be determined by the method described above as a method for measuring the ratio of the carbon atom content to the total content of all atoms at the center position in the film thickness direction of the support. Ten points are randomly selected from the center positions in the film thickness direction of the support determined by the above method, elemental analysis is performed, and the ratio of the silicon atom content to the total content of carbon atoms and silicon atoms at each point is calculated, and the average value of the 10 points is taken as the ratio X (atomic %) of the silicon atom content to the total content of carbon atoms and silicon atoms at the center position in the film thickness direction of the porous carbon support.

[0049] The contents of carbon atoms, silicon atoms, and metal atoms at a position 3 μm away from the interface between the porous carbon support and the dense carbon layer on the porous carbon support side can be determined by performing elemental analysis using an energy dispersive X-ray analyzer (EDX) at an acceleration voltage of 15 keV. 10 points are randomly selected from the positions 3 μm away from the interface between the support and the dense carbon layer, as determined by the method described below, and elemental analysis is performed to calculate the ratio of the content of silicon atoms to the total content of carbon atoms and silicon atoms at each point, and the ratio of the content of metal atoms to the total content of carbon atoms and metal atoms at each point, and the average value of the 10 points is the ratio Y (atomic%) of the content of silicon atoms to the total content of carbon atoms and silicon atoms at a position 3 μm away from the interface between the porous carbon support and the dense carbon layer on the porous carbon support side, and the ratio Z (atomic%) of the content of metal atoms to the total content of carbon atoms and metal atoms at a position 3 μm away from the interface between the porous carbon support and the dense carbon layer on the porous carbon support side.

[0050] The position on the porous carbon support side 3 μm from the interface between the porous carbon support and the dense carbon layer refers to the position obtained by moving the field of view 3 μm from the interface between the support and the dense carbon layer, determined by the above-mentioned method, toward the support side in a direction perpendicular to the interface with the dense carbon layer. This will be explained using Figure 4. If the destination of the 3 μm field of view is within the support, such as a point on the dense carbon layer, measurement is possible8. If the destination of the 3 μm field of view is a point outside the support, such as a point on the dense carbon layer, measurement is not possible9, and a different point is selected.

[0051] [Dense carbon layer] The dense carbon layer is formed on a support and functions as a separation layer for fluids. The dense carbon layer is made of carbon, and is a layer in which no pores are observed on the surface or cross section with a scanning electron microscope. "No pores observed with a scanning electron microscope" means that no clear pores are observed when a cross section formed by the CP method is observed at a magnification of 1±0.1 (nm / pixel). Therefore, it also includes those having only pores below the resolution. Here, "the dense carbon layer is made of carbon" means that the ratio of carbon atoms to the total number of atoms contained in the dense carbon layer is maximum.

[0052] The dense carbon layer is usually formed on the outer surface side of the support, but in the case of a hollow fiber carbon membrane for fluid separation, it may be formed on the inner surface side, i.e., the surface side in contact with the hollow portion, or it may be formed on both the outer surface side and the inner surface side. In the case of a film-shaped carbon membrane for fluid separation, the dense carbon layer may be formed on both sides of the film or on only one side. When the dense carbon layer is formed on both sides, the fluid can be separated by the dense carbon layers on both sides by supplying the fluid to the support from the cross section of the film.

[0053] The thickness of the dense carbon layer, i.e., the membrane thickness, is not particularly limited and can be appropriately set depending on the application, etc. In general, the thinner the membrane thickness, the higher the fluid permeation rate, so it is preferably 10 μm or less, more preferably 5 μm or less, and even more preferably 1 μm or less. On the other hand, the thicker the membrane thickness, the higher the separation function by suppressing fluid leakage, so it is preferably 1 nm or more, and more preferably 10 nm or more. Here, the membrane thickness of the dense carbon layer is the average value of the length of the line segment connecting two points, which are the shortest distance between the point on the interface between the support and the dense carbon layer determined by the above-mentioned method and the point on the surface of the other dense carbon layer other than the surface of the dense carbon layer that is the interface with the support, measured at 10 randomly selected points.

[0054] From the viewpoint of improving the separation factor, the ratio of silicon atom content to the total content of carbon atoms and silicon atoms at the central position in the thickness direction of the dense carbon layer is preferably 10 atomic % or less, more preferably 3 atomic % or less, and even more preferably 0.5 atomic % or less.

[0055] The carbon atom and silicon atom contents at the center position in the thickness direction of the dense carbon layer can be measured by elemental analysis using an energy dispersive X-ray analyzer (EDX) at an acceleration voltage of 5 keV. Elemental analysis is performed by randomly selecting 10 points from the center positions of the dense carbon layer thickness determined as described below, and the ratio of the silicon atom content to the total content of carbon atoms and silicon atoms at each point is calculated, and the average value of the 10 points is taken as the ratio of the silicon atom content to the total content of carbon atoms and silicon atoms at the center position in the thickness direction of the dense carbon layer.

[0056] The central position in the thickness direction of the dense carbon layer refers to the midpoint (point F in FIG. 5) of a line segment connecting a point (point D in FIG. 5) on the interface 7 between the support 5 and the dense carbon layer 6, which is determined by the above-mentioned method, and a point (point E in FIG. 5) on the other surface of the dense carbon layer other than the surface of the dense carbon layer that serves as the interface with the support, which is the shortest distance from the point on the interface.

[0057] <Method of manufacturing carbon membrane for fluid separation> Examples of the method for producing the carbon membrane for fluid separation of the present invention include a method having a step of carbonizing a molded body containing a resin that serves as a precursor of the support to obtain a support (step 1), a step of forming a carbonizable resin layer that serves as a precursor of a dense carbon layer on the support (step 2), and a step of carbonizing the carbonizable resin layer to form a dense carbon layer (step 3).

[0058] [Process 1] Step 1 is a step of obtaining a support by carbonizing a molded body containing a resin that serves as a precursor of the support (hereinafter, sometimes referred to as a "support precursor resin").

[0059] Examples of the support precursor resin include thermoplastic resins and thermosetting resins. Examples of the thermoplastic resin include polyphenylene ether, polyvinyl alcohol, polyacrylonitrile, phenolic resin, aromatic polyester, polyamic acid, aromatic polyimide, aromatic polyamide, polyvinylidene fluoride, cellulose acetate, polyetherimide, and copolymers thereof. Examples of the thermosetting resin include unsaturated polyester resin, alkyd resin, melamine resin, urea resin, polyimide resin, diallyl phthalate resin, lignin resin, urethane resin, phenolic resin, polyfurfuryl alcohol resin, and copolymers thereof. Two or more of these may be used. Among these, solution-spinnable thermoplastic resins are preferred, and polyacrylonitrile and aromatic polyimide are particularly preferred.

[0060] The weight average molecular weight (MW) of the support precursor resin is preferably 10,000 or more from the viewpoint of suppressing thread breakage in the spinning process and film rupture in the film forming process, while the weight average molecular weight (MW) of the support precursor resin is preferably 1,000,000 or less from the viewpoint of improving moldability.

[0061] In addition to the support precursor resin, it is preferable to add a disappearing component that can be disappeared after molding to the molded body containing the support precursor resin. For example, by preparing a resin mixture with a resin that disappears by subsequent heating such as during carbonization, or by dispersing particles that disappear by subsequent heating such as during carbonization, a porous structure can be formed and the average diameter of the pores that form the porous structure of the porous structure can be easily adjusted to a desired range.

[0062] As an example of a method for obtaining a molded product containing a support precursor resin, an example of adding a resin that disappears after carbonization (eliminating resin) will be described. First, the support precursor resin and the eliminating resin are mixed to obtain a resin mixture. The mixing ratio is preferably 10 to 90% by weight of the support precursor resin and 10 to 90% by weight of the eliminating resin.

[0063] Here, it is preferable to select a resin that is compatible with the carbonizable resin as the eliminator resin. The method of compatibility may be a mixture of the resins alone, or a solvent may be added. The combination of the carbonizable resin and the eliminator resin is not limited, but examples thereof include polyacrylonitrile / polyvinyl alcohol, polyacrylonitrile / polyvinyl phenol, polyacrylonitrile / polyvinyl pyrrolidone, and polyacrylonitrile / polylactic acid.

[0064] The obtained resin mixture is preferably subjected to phase separation during the molding process. The method for phase separation is not limited, and examples thereof include a thermally induced phase separation method and a non-solvent induced phase separation method.

[0065] When the support is fibrous, a precursor of the support can be formed by solution spinning. Solution spinning is a method in which a resin is dissolved in various solvents to prepare a spinning dope, and the dope is passed through a bath of a solvent that is a poor solvent for the resin to solidify the resin and obtain fibers. Examples of solution spinning include dry spinning, dry-wet spinning, and wet spinning.

[0066] For example, a hollow fiber-shaped molded product can be obtained by extruding the resin mixture prepared as described above from the outer tube of a hollow fiber spinneret having a double tube structure, and simultaneously discharging a gas such as air or nitrogen, the same solvent as the spinning dope, a solution in which the elimination resin has been dissolved, and the like from the inner tube of the spinneret.

[0067] In addition, by appropriately controlling the spinning conditions, it is possible to suppress the formation of a dense layer on the outer periphery of the fiber and open the surface of the porous carbon support. For example, when spinning is performed using a non-solvent induced phase separation method, the composition and temperature of the spinning dope and coagulation bath are appropriately controlled, or a spinning solution is discharged from an inner tube and a solution in which the same solvent as the spinning solution or a dissolving resin is dissolved is simultaneously discharged from an outer tube.

[0068] The fibers spun by the above-mentioned method are coagulated in a coagulation bath, followed by washing with water and drying to obtain a precursor of the support. Examples of the coagulation liquid include water, ethanol, saline, and mixtures of these with the solvent used in step 1. It is also possible to dissolve the solvent and the elimination resin by immersing the fibers in a coagulation bath or water bath before the drying step.

[0069] On the other hand, when the support is in the form of a film, a molded product containing the support precursor resin can be produced by a known casting method or spin coating method.

[0070] When the support is fibrous, it is preferable to apply the silicon-containing component after washing with water. When the support is film-like, it is preferable to apply the silicon-containing component after removing the solvent component. Examples of the silicon-containing component include powder of silicone oil or silicone resin, silica particles, etc. In order to improve dispersibility, the powder or particles may be surface-treated. In addition, an emulsifier may be applied together with the silicon-containing component. From the viewpoint of applying the silicon-containing component uniformly, it is preferable to use the silicon-containing component in the form of an emulsion, slurry, or solution. As the solvent for dissolving or dispersing the silicon-containing component, a poor solvent for the support precursor resin is preferable, and water, ethanol, acetone, etc. are preferable.

[0071] When spinning by the so-called multifilament method in which two or more fibrous support precursors are simultaneously processed in a contacted state, it is preferable to add a silicon-containing component before drying in order to prevent the fibers from fusing together when heated. By adding a silicon-containing component before drying, the occurrence of defects on the fiber surface, i.e., the support surface, when the fused surfaces of the fibers are peeled off can be further suppressed.

[0072] Examples of the method for applying the silicon-containing component include a dip method in which the support precursor is immersed in a liquid containing silicon, a nozzle coating method in which the support precursor is coated or sprayed with a liquid containing silicon, a spray coating method, and a cast method. In terms of ease of production, when the support precursor is in the form of a film, the dip method and the cast method are preferred, and when the support precursor is in the form of a fiber, the dip method and the nozzle coating method are preferred.

[0073] Examples of liquids containing silicon include a slurry in which silica particles are dispersed, an emulsion in which silicone oil is dispersed, etc. The degree of uneven distribution can be adjusted by adjusting the silicon atom concentration in the liquid.

[0074] When the silicon-containing component is applied by the dipping method, the dipping time is preferably 1 second or more, more preferably 5 seconds or more, and even more preferably 10 seconds or more, from the viewpoint of easily adjusting the value of Y to the above-mentioned preferred range, while the dipping time is preferably 2 minutes or less, more preferably 1 minute or less, and even more preferably 30 seconds or less, from the viewpoint of easily adjusting the value of Y to the above-mentioned preferred range.

[0075] The content of the silicon-containing component in the support precursor before infusibility treatment is preferably 0.1% or more, more preferably 0.5% or more, from the viewpoint of preventing fusion, and is preferably 50% or less, more preferably 30% or less, and even more preferably 15% or less, from the viewpoint of efficiently proceeding with the infusibility treatment and improving the permeability of the support.

[0076] The amount of silicon-containing component attached is calculated by taking about 3 g of the support precursor, weighing the precursor weight W1 after drying at 80°C for 24 hours, immersing the weighed support precursor in hexane for 6 hours to extract the silicon-containing component, removing the support precursor from the hexane, washing with pure water, and then drying at 80°C for 24 hours, and then measuring the weight W2 of the support precursor, and calculating it according to the following formula. (Amount of silicon-containing component attached) = (W1 - W2) / W2 × 100 [%].

[0077] When the support contains metal atoms, a metal-containing component may be added to a support precursor resin solution, or the metal-containing component may be applied by a dipping method, a nozzle coating method, etc. using the metal-containing component. The metal-containing component may be mixed with the silicon-containing component to simultaneously apply metal and silicon.

[0078] Examples of the metal-containing component include sodium chloride, calcium chloride, magnesium chloride, etc. The amount of metal atoms to be applied can be adjusted by the concentration of the metal atoms in the metal-containing component, the contact time between the metal-containing component and the support precursor, etc.

[0079] The precursor of the support prepared by the above-mentioned method can be subjected to an infusibility treatment before being subjected to a carbonization treatment. Examples of infusibility treatment include a method of heating the precursor in the presence of oxygen to cause oxidative crosslinking, a method of irradiating the precursor with high-energy rays such as electron beams and gamma rays to form a crosslinked structure, and a method of impregnating and mixing a substance having a reactive group into the precursor to form a crosslinked structure. Two or more of these methods may be combined. Among them, the method of heating the precursor in the presence of oxygen to cause oxidative crosslinking is preferred because the process is simple and the production cost is low.

[0080] The precursor of the support, which has been subjected to infusibility treatment as necessary, is finally carbonized to become the support. Carbonization is preferably performed by heating in an inert gas atmosphere. Examples of inert gas include helium, nitrogen, and argon. The flow rate of the inert gas may be an amount that can sufficiently reduce the oxygen concentration in the heating device, and it is preferable to select an appropriate value depending on the size of the heating device, the supply amount of the raw material, the carbonization temperature, etc. The lost resin may be removed by pyrolysis due to the heat during carbonization.

[0081] When the carbonization treatment is carried out continuously, in order to increase productivity, it is preferable to use a method in which the support is continuously fed into a heating device maintained at a constant temperature using a roller or conveyor and then removed.

[0082] When carbonization is performed by batch processing, the heating rate and the temperature decreasing rate can be set arbitrarily. From the viewpoint of productivity, the heating rate and the temperature decreasing rate are preferably 1° C. / min or more. On the other hand, there is no upper limit to the heating rate and the temperature decreasing rate, and they can be set arbitrarily within a range in which defects such as cracks do not occur.

[0083] The retention time at the carbonization temperature can be set arbitrarily. The retention time can be set within a range in which the support does not shrink in the second carbonization treatment (step 3) for forming a dense carbon layer described later, and is preferably 1 minute to 3 hours.

[0084] The carbonization temperature is preferably 500°C or higher and 2,400°C or lower. Here, the carbonization temperature is the maximum temperature reached when performing the carbonization treatment. From the viewpoint of improving the function as a support, the carbonization temperature is more preferably 900°C or higher. On the other hand, from the viewpoint of reducing brittleness and improving handleability, the carbonization temperature is more preferably 1,500°C or lower.

[0085] [Process 2] Step 2 is a step of forming a carbonizable resin layer, which is a precursor of the dense carbon layer, on the support prepared in step 1. By preparing the support and the dense carbon layer in separate steps, the thickness of the dense carbon layer can be set as desired. This makes it easier to design the separation membrane structure, for example by reducing the thickness of the dense carbon layer to improve the fluid permeation rate.

[0086] As the carbonizable resin, various resins that exhibit fluid separation properties after carbonization can be adopted. Examples of the carbonizable resin include polyacrylonitrile, aromatic polyimide, polybenzoxazole, aromatic polyamide, polyphenylene ether, phenol resin, cellulose acetate, polyfurfuryl alcohol, polyvinylidene fluoride, lignin, wood tar, and polymer of inherent porosity (PIM). Two or more of these may be used. Among these, polyacrylonitrile, aromatic polyimide, polybenzoxazole, aromatic polyamide, polyphenylene ether, and polymer of inherent porosity (PIM) are preferable because they have excellent fluid permeation speed and separation properties, and polyacrylonitrile and aromatic polyimide are more preferable. The carbonizable resin may be the same as or different from the above-mentioned support precursor resin.

[0087] The method for forming the carbonizable resin layer is not limited, and known methods can be used. A typical method is to coat the carbonizable resin itself on the support, but other methods include a method in which a precursor of the resin is coated on the support and then the precursor is reacted to form the carbonizable resin layer, and a counter diffusion method in which reactive gases or solutions are flowed from the outside and inside of the support to cause a reaction. Examples of the reaction include polymerization, cyclization, and crosslinking reactions caused by heating or a catalyst.

[0088] Examples of the coating method for the carbonizable resin layer include a dip coating method, a nozzle coating method, a spray method, a vapor deposition method, a cast coating method, etc. From the viewpoint of ease of production, when the support is fibrous, the dip coating method or the nozzle coating method is preferred, and when the support is film-shaped, the dip coating method or the cast coating method is preferred.

[0089] The dip coating method is a method in which a support is immersed in a coating liquid containing a solution of a carbonizable resin or its precursor, and then withdrawn.

[0090] The viscosity of the coating solution in the dip coating method can be set arbitrarily depending on conditions such as the surface roughness of the support, the lifting speed, and the desired film thickness. If the viscosity of the coating solution is high, a uniform resin layer can be formed. Therefore, at a shear rate of 0.1 s -1 The shear viscosity at a shear rate of 0.1 s is preferably 10 mPa·s or more, and more preferably 50 mPa·s or more. On the other hand, the lower the viscosity of the coating solution, the thinner the film becomes and the higher the fluid permeation rate. -1 The shear viscosity at is preferably 1,000 mPa·s or less, and more preferably 800 mPa·s or less.

[0091] The lifting speed of the support in the dip coating method can also be set arbitrarily depending on the coating conditions. If the lifting speed is fast, the thickness of the carbonizable resin layer will be thicker, and defects in the carbon film can be suppressed. Therefore, the lifting speed is preferably 1 mm / min or more, more preferably 10 mm / min or more. On the other hand, if the lifting speed is slow, the film thickness uniformity of the carbonizable resin layer will be improved, and the permeation speed of the fluid will be improved. Therefore, the lifting speed is preferably 1,000 mm / min or less, more preferably 800 mm / min or less. The temperature of the coating stock solution is preferably 20°C or more and 80°C or less. If the temperature of the coating stock solution is high, the surface tension will be reduced, the wettability to the support will be improved, and the thickness of the carbonizable resin layer will be uniform.

[0092] The nozzle coating method is a method for laminating a resin or a resin precursor on a porous carbon support by passing the porous carbon support through a nozzle filled with a coating liquid, which is a solution of a carbonizable resin or its precursor. The viscosity and temperature of the coating liquid, the nozzle diameter, and the passing speed of the porous carbon support can be set arbitrarily.

[0093] When treating the multifilament in the subsequent infusible treatment step or carbonization step (step 3), it is preferable to apply an oil agent to suppress the formation of defects during fusion and splitting. Examples of the oil agent include silicone-based oil agents and non-silicone-based oil agents. Examples of the silicone-based oil agent include amino-modified silicone and epoxy-modified silicone. Two or more of these may be used.

[0094] [Infusible treatment] The support having a carbonizable resin layer formed thereon, prepared in step 2 (hereinafter referred to as "support / carbonizable resin layer composite"), may be subjected to an infusible treatment before the carbonization treatment (step 3). The method of the infusible treatment is not limited, and is similar to the infusible treatment of the precursor of the porous carbon support described above.

[0095] [Step 3] Step 3 is a step of heating the support / carbonizable resin layer composite produced in step 2 and further subjected to an infusible treatment as necessary, to carbonize the carbonizable resin layer and form a dense carbon layer.

[0096] In this step, it is preferable to heat the support / carbonizable resin layer composite in an inert gas atmosphere. The flow rate of the inert gas may be any amount that can sufficiently reduce the oxygen concentration in the heating device, and it is preferable to select an appropriate optimum value depending on the size of the heating device, the supply amount of the raw material, the carbonization temperature, etc. There is no upper limit to the flow rate of the inert gas, but it is preferable to set it appropriately in accordance with the temperature distribution and the design of the heating device from the viewpoint of economy and minimizing temperature changes in the heating device.

[0097] In addition, by heating in a mixed gas atmosphere of the above-mentioned inert gas and active gas, the surface of the porous carbon support can be chemically etched, and the pore diameter of the support surface can be adjusted to a desired range. Examples of active gases include oxygen, carbon dioxide, water vapor, air, and combustion gas. The concentration of the active gas in the inert gas is preferably 0.1 ppm or more and 100 ppm or less.

[0098] The carbonization temperature in this step can be set arbitrarily within a range in which the permeation rate and separation factor of the carbon membrane for fluid separation are improved, but it is preferable that the carbonization temperature is lower than the carbonization temperature when the precursor of the support is carbonized in step 1. This makes it possible to reduce the moisture absorption dimensional change rate of the support and the carbon membrane for fluid separation, thereby suppressing breakage of the carbon membrane for fluid separation in the separation module, while improving the permeation rate and separation performance of the fluid. The carbonization temperature in this step is preferably 500°C or higher, more preferably 550°C or higher. Also, it is preferably 850°C or lower, more preferably 800°C or lower.

[0099] Other preferable aspects of the carbonization are the same as those of the carbonization of the precursor of the support described above.

[0100] [Post-processing process] The carbon membrane for fluid separation produced by steps 1 to 3 can be subjected to various known post-treatments in order to obtain a desired permeation rate and separation factor. Examples of the post-treatments include heat treatment and pore control by a chemical vapor deposition (CVD) method. EXAMPLES

[0101] Preferred examples of the present invention will be described below, but the present invention is not limited to these examples. 4、5、8 is a comparative example 4、5、8 First, the evaluation method in each example will be described.

[0102] (Measurement of silicon atom, carbon atom, and metal atom content) For the carbon membranes for fluid separation obtained in each of the Examples and Comparative Examples, the surface formed by the CP method was observed at a magnification of 1±0.1 (nm / pixel) while moving the field of view from the dense carbon layer to the porous carbon support side, and the edge of the first observed pore on the dense carbon layer side was determined to be the interface between the dense carbon layer and the porous carbon support.

[0103] At ten points randomly selected from the center positions in the film thickness direction of the support, that is, the midpoints of the line segments connecting a point on the interface between the dense carbon layer and the support and a point on the surface of the other support that has the smallest length of the line segment connecting the point on the interface, elemental analysis was performed using an energy dispersive X-ray analyzer (S-5500, manufactured by Hitachi High-Technologies Corporation) at an acceleration voltage of 15 keV to measure the contents of carbon atoms, silicon atoms, and metal atoms. The ratio X of the silicon atom content to the total content of carbon atoms and silicon atoms at the center position in the film thickness direction of the porous carbon support was calculated from the average value of the ten points.

[0104] The field of view was shifted 3 μm from the determined interface toward the porous membrane in a direction perpendicular to the surface of the dense carbon layer, and elemental analysis was performed at 10 randomly selected locations 3 μm away from the interface using an energy dispersive X-ray analyzer (S-5500, manufactured by Hitachi High-Technologies Corporation) at an acceleration voltage of 15 keV to measure the content of each element, carbon atom, silicon atom, and metal atom. From the average value of the 10 locations, the ratio Y of the content of silicon atoms to the total content of carbon atoms and silicon atoms at a position 3 μm away from the interface between the porous carbon support and the dense carbon layer on the porous carbon support side, and the ratio Z of the content of metal atoms to the total content of carbon atoms and metal atoms at a position 3 μm away from the interface between the porous carbon support and the dense carbon layer on the porous carbon support side were calculated.

[0105] (Measurement of the number of defects) The carbon membranes for fluid separation obtained in each of the Examples and Comparative Examples were cut to a length of 20 cm to prepare three samples, one end of each membrane was sealed, and the three membranes were simultaneously left in the air for 10 seconds while compressed air was supplied to the hollow parts from the other end at 0.2 MPaG. Then, the membranes were immersed in water while still under pressure and left to stand for 5 seconds. After 5 seconds, the number of bubbles adhering to the surface of each carbon membrane was visually counted, and the total number of bubbles was taken as the number of defects.

[0106] (Measurement of separation factor) For the carbon membranes for fluid separation obtained in each of the Examples and Comparative Examples, modules were fabricated using the carbon membranes for fluid separation in which the number of defects was measured and the defects were repaired, and the gas permeation rates were measured. Carbon dioxide and methane were used as the measurement gases, and the pressure changes per unit time of the permeation side of carbon dioxide and methane were measured by an external pressure method at a measurement temperature of 25°C in accordance with the pressure sensor method of JIS K7126-1 (2006). Here, the pressure difference between the supply side and the permeation side was set to 0.11 MPa.

[0107] Next, the gas permeation rate Q was calculated by the following formula. The ratio of the gas permeation rates of each component was defined as the separation factor α. Here, the membrane area was calculated from the outer diameter and length of the region that contributes to gas permeation. The repaired defect portion was not included in the membrane area because it did not contribute to permeation.

[0108] Permeation rate Q = [permeation gas amount (mol)] / [membrane area (m 2 ) × time (s) × pressure difference (Pa) The permeation rates Q of carbon dioxide and methane were measured, and the separation factor was calculated as (permeation rate of carbon dioxide) / (permeation rate of methane).

[0109] [Example 1] 10 parts by weight of polyacrylonitrile (PAN) (MW 150,000) manufactured by Polysciences Corporation, 10 parts by weight of polyvinylpyrrolidone (PVP) (MW 40,000) manufactured by Sigma-Aldrich Corporation, and 80 parts by weight of dimethyl sulfoxide (DMSO) manufactured by Fujifilm Wako Pure Chemical Industries, Ltd. were mixed and stirred at 100°C to prepare a spinning dope.

[0110] After cooling the obtained spinning solution to 25°C, a concentric triple die was used to simultaneously discharge an 80% by weight aqueous solution of DMSO from the inner tube, the spinning solution from the middle tube, and a 90% by weight aqueous solution of DMSO from the outer tube, and then the spinning solution was introduced into a coagulation bath of pure water at 25°C and wound up on a roller to obtain a raw yarn. Next, both ends of the raw yarn were sealed, and the yarn was dipped for 20 seconds into an emulsion of epoxy-modified silicone oil manufactured by Shin-Etsu Chemical Co., Ltd. dispersed in water using an emulsifier containing sodium to impart silicon, and dried at 80°C for 1 hour using a circulating hot air dryer to produce a hollow fiber-shaped porous carbon support precursor.

[0111] The porous carbon support precursor was then passed through an electric furnace at 260°C and heated in an air atmosphere for 1 hour to perform an infusible treatment. The infusible fibers were then carbonized at a carbonization temperature of 650°C to produce a hollow fiber-shaped porous carbon support. Both the outer and inner surfaces (surfaces of the hollow portion) of the produced porous carbon support were open, and when the cross section of the hollow fiber was observed, a bicontinuous porous structure was observed.

[0112] Next, the spinning dope was applied as a coating liquid to the support surface by a nozzle coating method to form a dense carbon layer precursor coating, and a carbon membrane precursor for fluid separation was obtained. The obtained carbon membrane precursor was passed through an electric furnace at 260°C and heated in an air atmosphere for 1 hour to perform infusibility treatment. The carbon membrane precursor that had been infusible treated was then carbonized at a carbonization temperature of 650°C to produce a hollow fiber carbon membrane for fluid separation.

[0113] The carbon membrane for fluid separation thus produced was measured for X and Y by the above-mentioned method, and found to be 3.1 atomic % for X and 8.9 atomic % for Y. Sodium was detected as a metal atom on the support side, and Z was 0.023 atomic %. No defects were observed, and the carbon dioxide permeation rate Q was 0.23×10 -9 mol / (m 2 ·Pa·s), and the separation factor was 42.1.

[0114] [Example 2] A carbon membrane for fluid separation was prepared in the same manner as in Example 1, except that the immersion time in the emulsion was set to 10 seconds.

[0115] [Example 3] A carbon membrane for fluid separation was prepared in the same manner as in Example 1, except that the immersion time in the emulsion was changed to 5 seconds. The prepared carbon membrane for fluid separation was evaluated in the same manner as in Example 1.

[0116] [Example 4] A carbon membrane for fluid separation was produced in the same manner as in Example 1, except that silicon was applied by nozzle coating. The produced carbon membrane for fluid separation was evaluated in the same manner as in Example 1.

[0117] [Example 5] A carbon membrane for fluid separation was prepared in the same manner as in Example 1, except that the emulsion in which the epoxy-modified silicone oil was dispersed was replaced with epoxy-modified silicone oil manufactured by Shin-Etsu Chemical Co., Ltd., and the membrane was dipped for 5 seconds in the epoxy-modified silicone oil manufactured by Shin-Etsu Chemical Co., Ltd. The prepared carbon membrane for fluid separation was evaluated in the same manner as in Example 1.

[0118] [Example 6] A carbon membrane for fluid separation was prepared in the same manner as in Example 1, except that the emulsion in which epoxy-modified silicone oil was dispersed was replaced by a slurry in which nanosilica particles "Snowtex" (registered trademark) ZL manufactured by Nissan Chemical Industries, Ltd. were dispersed in water at a concentration of 5% by weight, and the membrane was dipped for 20 seconds in this slurry. The prepared carbon membrane for fluid separation was evaluated in the same manner as in Example 1.

[0119] [Example 7] A carbon membrane for fluid separation was prepared in the same manner as in Example 6, except that the membrane was dipped for 20 seconds into a slurry in which nanosilica particles "Snowtex" ZL manufactured by Nissan Chemical Industries, Ltd. were dispersed in water at a concentration of 4.7% by weight and magnesium chloride hexahydrate manufactured by Wako Pure Chemical Industries, Ltd. was dispersed in water at a concentration of 0.3% by weight, instead of the slurry in which nanosilica particles "Snowtex" ZL manufactured by Nissan Chemical Industries, Ltd. were dispersed in water at a concentration of 5% by weight. The prepared carbon membrane for fluid separation was evaluated in the same manner as in Example 1.

[0120] [Example 8] A carbon membrane for fluid separation was prepared in the same manner as in Example 1, except that nanosilica particles "Snowtex" ZL manufactured by Nissan Chemical Industries, Ltd. were added to the coating solution to a concentration of 1% by weight. The prepared carbon membrane for fluid separation was evaluated in the same manner as in Example 1.

[0121] [Comparative Example 1] A carbon membrane for fluid separation was prepared in the same manner as in Example 1, except that silicon was not added. The prepared carbon membrane for fluid separation was evaluated in the same manner as in Example 1.

[0122] [Comparative Example 2] A carbon membrane for fluid separation was produced in the same manner as in Example 1 except that the coating stock solution was prepared by mixing 24 parts by weight of polyethersulfone powder grade 5200P manufactured by Sumitomo Chemical Co., Ltd., 1 part by weight of magnesium chloride hexahydrate manufactured by Wako Pure Chemical Industries, Ltd., and 75 parts by weight of dimethylformamide manufactured by Wako Pure Chemical Industries, Ltd., and silicon was not added to the support precursor. The produced carbon membrane for fluid separation was evaluated in the same manner as in Example 1.

[0123] [Comparative Example 3] Using a double tube die, the spinning dope was simultaneously discharged from the outer tube and an 80 wt% DMSO aqueous solution was simultaneously discharged from the inner tube, and then immediately introduced into a coagulation bath of pure water at 25°C and wound on a roller to obtain an asymmetric hollow fiber raw fiber. The obtained fiber was washed with water at 25°C for 5 minutes and dried at 80°C for 1 hour using a circulating hot air dryer to obtain a hollow fiber carbon membrane precursor for fluid separation. The fluid separation carbon membrane precursor was then passed through an electric furnace at 260°C and heated in an air atmosphere for 1 hour to perform an infusible treatment. The infusible fiber was then carbonized at a carbonization temperature of 650°C to produce a hollow fiber carbon membrane for fluid separation. The produced carbon membrane for fluid separation was an asymmetric hollow carbon membrane, and a dense carbon layer was formed on the surface of a support layer made of porous carbon. The porous portion had a bicontinuous structure. The obtained carbon membrane for fluid separation was evaluated in the same manner as in Example 1.

[0124] The evaluation results of each of the examples and comparative examples are shown in Table 1.

[0125] [Table 1] [Explanation of symbols]

[0126] 1. Porous carbon support 2 Dense carbon layer 3 Branch 4 Cavity 5 Dense carbon layer 6 Support 7 Interface 8 measurable 9 Not measurable

Claims

1. A carbon membrane for fluid separation, comprising a dense carbon layer formed on a porous carbon support, the membrane thickness of the porous carbon support being 10 μm or more, and wherein, when a ratio of a content of silicon atoms to a total content of carbon atoms and silicon atoms at a central position in a membrane thickness direction of the porous carbon support is X (atomic %), and a ratio of a content of silicon atoms to a total content of carbon atoms and silicon atoms at a position 3 μm away from an interface of the porous carbon support with the dense carbon layer on the porous carbon support side is Y (atomic %), X<Y, a difference (Y-X) between Y and X is 0.10 or more and 10 or less, and the ratio of a content of silicon atoms to a total content of carbon atoms and silicon atoms at a central position in a membrane thickness direction of the dense carbon layer of the dense carbon layer is 0.5 atomic % or less.

2. 2. The carbon membrane for fluid separation according to claim 1, wherein the ratio of the content of metal atoms to the total content of carbon atoms and metal atoms at a position 3 μm away from the interface between the porous carbon support and the dense carbon layer on the porous carbon support side is defined as Z (atomic %), and Z exceeds 0.

3. The carbon membrane for fluid separation according to claim 1 or 2, wherein Z is 0.0010 or more.

4. 4. The carbon membrane for fluid separation according to claim 2, wherein the metal atoms are sodium and / or calcium.

Images