Carbon film composite and production method for carbon film composite

Abstract

A carbon film composite according to the present invention comprises a porous support (11) and a carbon film (12) that is provided on the surface of the support (11). A composite layer (13) that spreads into the support (11) from the surface of the support (11) is present in the support (11). At the composite layer (13), carbon and / or a carbide is embedded in gaps (333) between aggregate particles (332) of the support (11). The arithmetic mean of the thickness (i.e., the average thickness tave) of the composite layer (13) in a thickness direction that is orthogonal to the surface of the support (11) is 0.01–1.0 μm, and the standard deviation tsd of the thickness of the composite layer (13) is 0.1–1.0 μm. In other words, the thickness and the variation in the thickness of the composite layer (13) are reduced, and the carbon film composite can thereby increase permeation flux.

Description

Carbon membrane composite and method for producing the carbon membrane composite

[0001] The present invention relates to a carbon membrane composite and a method for manufacturing the same. [Reference to Related Applications] This application claims the benefit of priority from Japanese Patent Application JP2024-052164, filed on March 27, 2024, the entire disclosure of which is incorporated herein by reference.

[0002] Carbon membrane composites, which have a carbon membrane provided on the surface of a porous support, have been used as filters with excellent chemical stability and heat resistance. In these carbon membrane composites, it is required to make the carbon membrane thin and uniform in thickness in order to achieve both an increase in permeation flux and an improvement in separation performance.

[0003] For example, International Publication No. 2008 / 010452 (Document 1) proposes providing a surface deposition layer on the surface of a support such that the average pore diameter gradually decreases. The outermost layer of the surface deposition layer is formed from titania sol, and the average pore diameter is set to 0.3 nm to 20 nm. This prevents the precursor solution of the carbon membrane from penetrating into the support when producing a carbon membrane composite, and allows the precursor solution to be applied thinly and uniformly onto the support. As a result, the carbon membrane in the carbon membrane composite can have a thin and uniform thickness.

[0004] Furthermore, WO 2013 / 042262 (Document 2) proposes immersing a support in a suspension of phenolic resin when producing a carbon membrane composite using phenolic resin as a precursor of the carbon membrane. This suspension has a higher viscosity than a solution of phenolic resin, and therefore, it becomes possible to uniformly coat the support with phenolic resin with fewer coating times.

[0005] In the carbon membrane composite, a composite layer in which the support and the carbon membrane are present in combination is formed at the interface between the support and the carbon membrane due to the carbon membrane precursor solution that permeates from the surface of the support into the inside of the support during production.

[0006] In the manufacturing method of Document 1, the average pore size on the surface of the support is quite small, and the amount of precursor solution that penetrates into the support may be uneven depending on the position on the support surface. As a result, the thickness of the composite layer varies greatly, and there is a concern that the permeation flux may decrease due to pressure loss in the composite layer. In addition, in the manufacturing method of Document 2, the amount of precursor solution that penetrates into the support increases, resulting in a thick composite layer, and there is a concern that the permeation flux may decrease due to pressure loss in the composite layer.

[0007] The present invention is directed to a carbon membrane composite and aims to increase the permeation flux of the carbon membrane composite.

[0008] A first aspect of the present invention is a carbon membrane composite comprising a porous support and a carbon membrane provided on the surface of the support. A composite layer is provided inside the support, the composite layer extending from the surface of the support toward the interior of the support and having voids between aggregate particles of the support filled with carbon or carbide. In a thickness direction perpendicular to the surface of the support, the arithmetic mean thickness of the composite layer is 0.01 μm or more and 1.0 μm or less, and the standard deviation of the thickness of the composite layer is 0.1 μm or more and 1.0 μm or less.

[0009] This can increase the permeation flux of the carbon membrane composite.

[0010] A second aspect of the invention is a carbon membrane composite comprising a porous support and a carbon membrane provided on the surface of the support. A composite layer is provided inside the support, the composite layer extending from the surface of the support toward the interior of the support, and voids between aggregate particles of the support being filled with carbon or carbide. In a thickness direction perpendicular to the surface of the support, a lower region of the composite layer is defined as a region extending from the surface of the support to a distance four times the arithmetic mean of the thickness of the composite layer, excluding a region extending from the surface of the support to a distance equal to the arithmetic mean of the thickness of the composite layer. The porosity of the lower region of the composite layer is 5% or more and 25% or less.

[0011] A third aspect of the invention is the carbon membrane composite of the second aspect, wherein the ratio of carbon to aluminum in a region extending from the surface of the support to a depth four times the arithmetic mean of the thickness of the composite layer in the thickness direction is 0.05 or more and 0.2 or less.

[0012] A fourth aspect of the invention is the carbon membrane composite according to any one of the first to third aspects, wherein (D50-D10) / (D90-D10) in the thickness distribution of the composite layer is 0.2 or more and 0.4 or less.

[0013] A fifth aspect of the invention is the carbon membrane composite according to any one of the first to third aspects (or any one of the first to fourth aspects), wherein the average pore size on the surface of the support is 0.01 μm or more.

[0014] A sixth aspect of the present invention is the carbon membrane composite of the fifth aspect, wherein the average pore size on the surface of the support is 0.07 μm or more.

[0015] A seventh aspect of the invention is the carbon membrane composite of any one of aspects 1 to 3 (or any one of aspects 1 to 6), wherein the support includes a first layer extending from the surface of the support toward the inside of the support, a second layer extending from the first layer toward the inside of the support and having a larger average pore diameter than the first layer, and a third layer extending from the second layer toward the inside of the support and having a larger average pore diameter than the second layer. The average pore diameter of the third layer is 3 μm or more.

[0016] An eighth aspect of the invention is the carbon membrane composite according to any one of the first to third aspects (or any one of the first to seventh aspects), wherein the thickness of the carbon membrane in the thickness direction is 0.1 μm or more and 5 μm or less.

[0017] A ninth aspect of the invention is the carbon membrane composite of any one of aspects 1 to 3 (or any one of aspects 1 to 8), wherein the surface roughness of the carbon membrane is 100 nm or more and 300 nm or less.

[0018] A tenth aspect of the invention is a method for producing a carbon membrane composite, comprising: a) applying a non-solvent to a porous support to allow the non-solvent to penetrate the support, thereby forming a non-solvent layer inside the support; b) removing a part of the non-solvent layer from the surface side of the support; c) applying a precursor solution containing a precursor from the surface side of the support while applying pressure from the side opposite to the surface of the support, thereby forming a precursor film, which is a film containing the precursor, on the surface of the support; d) drying the precursor film and the non-solvent layer while applying pressure from the side opposite to the surface of the support; and e) carbonizing the precursor film to form a carbon film on the surface of the support.

[0019] An eleventh aspect of the invention is a method for producing a carbon membrane composite, comprising: a) applying a liquid having precursor particles dispersed therein from the surface side of the support to form a layer of precursor particles on the surface of the support; b) melting the layer of precursor particles to form a precursor film, which is a film containing the precursor, on the surface of the support; and c) carbonizing the precursor film to form a carbon film on the surface of the support.

[0020] The above and other objects, features, aspects and advantages will become more apparent from the following detailed description of the invention which proceeds with reference to the accompanying drawings.

[0021] 1 is a side view of a carbon membrane composite according to one embodiment; FIG. 2 is a longitudinal cross-sectional view of a carbon membrane composite; FIG. 3 is a cross-sectional view showing an enlarged portion of the carbon membrane composite; FIG. 4 is a cross-sectional view showing an enlarged portion of the carbon membrane composite; FIG. 5 is a cross-sectional view showing an enlarged portion of the carbon membrane composite; FIG. 6 is a diagram showing the thickness distribution of a composite layer; FIG. 7 is a cross-sectional view showing an enlarged portion of a carbon membrane composite; FIG. 8 is a diagram showing the results of elemental mapping analysis of a polished cross-section; FIG. 9 is a diagram showing the flow of production of a carbon membrane composite; FIG. 10 is a cross-sectional view of a carbon membrane composite in the process of production; FIG. 11 is a cross-sectional view of a carbon membrane composite in the process of production; FIG. 12 is a cross-sectional view showing a separation device; FIG. 13 is a diagram showing the flow of production of a carbon membrane composite; FIG. 14 is a cross-sectional view of a carbon membrane composite in the process of production; FIG. 15 is a cross-sectional view of a carbon membrane composite in the process of production; FIG. 16 is a cross-sectional view of a carbon membrane composite in the process of production;

[0022] Fig. 1 is a side view showing an end face of a carbon membrane composite 1 according to one embodiment of the present invention. Fig. 2 is a longitudinal cross-sectional view of the carbon membrane composite 1. The carbon membrane composite 1 is a substantially columnar member extending in the left-right direction in Fig. 2. In the following description, the left-right direction in Fig. 2 will also be referred to as the "longitudinal direction." Fig. 1 shows an end face on one side in the longitudinal direction of the carbon membrane composite 1, but the end face on the other side has the same shape as that shown in Fig. 1.

[0023] The carbon membrane composite 1 includes a support 11 and a carbon membrane 12. The support 11 is a porous member that is permeable to gases and liquids. The support 11 illustrated in FIGS. 1 and 2 has a monolithic structure in which a single, continuous columnar body is provided with a plurality of through-holes 111 (hereinafter also referred to as "cells 111") that extend in the longitudinal direction of the body. The monolithic structure is a concept that includes a honeycomb structure. The outer shape of the support 11 is, for example, approximately cylindrical. The multiple cells 111 are arranged, for example, in a substantially concentric pattern in a cross section of the support 11 perpendicular to the longitudinal direction. The cross section of each cell 111 perpendicular to the longitudinal direction has, for example, an approximately circular shape. Note that "approximately circular" includes not only a perfect circle but also an ellipse or a distorted circle. In FIG. 2, the diameter of the cells 111 is larger than the actual diameter, and the number of cells 111 is smaller than the actual number. Note that the number, cross-sectional shape, and arrangement of the cells 111 are not limited to the above example and may be variously modified.

[0024] The length of the support 11 (i.e., the length in the left-right direction in FIG. 2) is, for example, 10 cm to 200 cm. The outer diameter of the support 11 is, for example, 0.5 cm to 30 cm. The distance between the central axes of adjacent cells 111 is, for example, 0.3 mm to 10 mm. The inner diameter of the cells 111 is, for example, 0.5 mm to 25 mm. The shape of the support 11 may be, for example, a honeycomb shape, a flat plate shape, a tubular shape, a cylindrical shape, a columnar shape, or a polygonal columnar shape. When the shape of the support 11 is tubular or cylindrical, the thickness of the support 11 is, for example, 0.1 mm to 10 mm.

[0025] A carbon film 12 is provided on the inner surface of each cell 111. The carbon film 12 is preferably provided so as to cover substantially the entire inner surface of each cell 111. The support 11 is used to support the carbon film 12. In Figures 1 and 2, the carbon film 12 is depicted by a thick line.

[0026] Fig. 3 is an enlarged cross-sectional view of a portion of the carbon membrane composite 1 shown in Fig. 2. In Fig. 3, the carbon membrane 12 is shown with hatched lines to make the thickness of the carbon membrane 12 appear thicker than it actually is. In addition, the diameter of the cells 111 is shown smaller than it actually is. In Fig. 3, the particles constituting the support 11 are schematically depicted as circles, but the particles may have various shapes.

[0027] In the example shown in FIG. 3 , the support 11 includes a substrate 31, an intermediate layer 32, and a surface layer 33. The substrate 31, the intermediate layer 32, and the surface layer 33 are each porous. The substrate 31 has the monolith structure described above. In the support 11, the intermediate layer 32 and the surface layer 33 are provided on the inner surface of each through-hole of the substrate 31 (i.e., the through-hole that will become the cell 111), thereby forming the cell 111. The intermediate layer 32 is provided on the inner surface of each through-hole of the substrate 31 so as to be in direct contact with the inner surface and cover substantially the entire inner surface. The intermediate layer 32 is a substantially cylindrical portion that is relatively thin in thickness in the radial direction (hereinafter simply referred to as the "radial direction") about the central axis of the cell 111.

[0028] The surface layer 33 is provided on the inner surface of the intermediate layer 32 so as to be in direct contact with the inner surface and to cover substantially the entire inner surface. That is, the surface layer 33 is provided indirectly on the substrate 31 via the intermediate layer 32. The surface layer 33 is a substantially cylindrical portion with a relatively thin radial thickness. When viewed from the inside of the cells 111, the surface layer 33 is the innermost portion of the support 11. That is, the inner surface of the surface layer 33 is the inner surface of the cells 111. A carbon film 12 is provided on the inner surface of the surface layer 33 so as to cover substantially the entire inner surface. The carbon film 12 is in direct contact with an inner surface 331 of the surface layer 33. That is, the inner surface of the surface layer 33 is the surface of the support 11 on which the carbon film 12 is formed.

[0029] The region near the inner surface of the cell 111 has a laminated structure in which the surface layer 33, the intermediate layer 32, and the substrate 31 are laminated in the thickness direction (i.e., in the radial direction about the central axis of the cell 111). In this laminated structure, the surface layer 33 is a first layer that extends from the surface of the support 11 toward the inside of the support 11 (i.e., from the inner surface of the cell 111 toward the radially outward direction). The intermediate layer 32 is a second layer that extends from the surface layer 33 toward the inside of the support 11 and is provided between the surface layer 33 and the substrate 31. The substrate 31 is a third layer that extends from the intermediate layer 32 toward the inside of the support 11.

[0030] The thickness of the surface layer 33 in the thickness direction (i.e., the direction perpendicular to the surface of the support 11) is, for example, 1 μm or more and 100 μm or less. The thickness of the intermediate layer 32 in the thickness direction is, for example, 100 μm or more and 500 μm or less. The thickness of the surface layer 33 in the thickness direction (hereinafter simply referred to as "thickness") can be determined, for example, by observing the polished cross section of the sample with a scanning electron microscope (SEM) at a magnification of 5000 times or more. The thickness of the intermediate layer 32 can also be determined similarly, for example, by observing the polished cross section of the sample with an SEM.

[0031] In this embodiment, the material of the support 11 (i.e., the base material 31, the intermediate layer 32, and the surface layer 33) is a ceramic that is chemically stable during the process of forming the carbon film 12. The support 11 is formed, for example, of a ceramic sintered body. Examples of ceramic sintered bodies selected as the material of the support 11 include alumina, silica, mullite, zirconia, titania, yttria, silicon nitride, and silicon carbide. Preferably, the support 11 is formed of alumina, mullite, zirconia, or titania. In this embodiment, alumina is used as the material of the support 11 in consideration of its high corrosion resistance, small change in pore size due to temperature changes, and relatively high strength. The materials of the base material 31, the intermediate layer 32, and the surface layer 33 may be the same or different.

[0032] The support 11 (i.e., the base material 31, the intermediate layer 32, and the surface layer 33) contains, for example, an inorganic binder for binding aggregate particles of the ceramic sintered body. The inorganic binder can be at least one of titania, mullite, sinterable alumina, silica, glass frit, clay minerals, and sinterable cordierite.

[0033] The support 11 may also contain an alkali metal and / or alkaline earth metal, such as sodium (Na), potassium (K), calcium (Ca), or magnesium (Mg).

[0034] The average pore diameter of the surface layer 33 is smaller than the average pore diameter of the intermediate layer 32 and the average pore diameter of the substrate 31. The average pore diameter of the intermediate layer 32 is smaller than the average pore diameter of the substrate 31. In other words, the average pore diameter of the substrate 31 is larger than the average pore diameter of the intermediate layer 32, and the average pore diameter of the intermediate layer 32 is larger than the average pore diameter of the surface layer 33. The average pore diameter of the substrate 31 is, for example, 0.01 μm or more and 70 μm or less, preferably 0.05 μm or more and 25 μm or less. More preferably, the average pore diameter of the substrate 31 is 3 μm or more. The average pore diameter of the intermediate layer 32 is, for example, 0.01 μm or more and 10 μm or less. The average pore diameter of the surface layer 33 is, for example, 0.005 μm or more and 2.0 μm or less, preferably 0.05 μm or more and 0.5 μm or less. The average pore diameters of the substrate 31, intermediate layer 32 and surface layer 33 can be measured by, for example, a mercury porosimeter, a perm porometer or a nanoperm porometer.

[0035] The porosity of the surface layer 33, the intermediate layer 32, and the substrate 31 is, for example, approximately the same. The porosity of the surface layer 33, the intermediate layer 32, and the substrate 31 may be different. The porosity of the surface layer 33, the intermediate layer 32, and the substrate 31 is, for example, 20% or more and 50% or less.

[0036] The porosity of the substrate 31 can be determined by the following procedure. First, the pores of the substrate 31 are filled with resin, and a polished cross section is prepared by mechanical polishing. Next, the polished cross section is observed using an SEM to obtain an image. The image is then binarized, and the pore portion, particle portion, and inorganic binder portion are color-coded, and the proportion of the pore portion to the entire surface is determined as the porosity. The porosity of the intermediate layer 32 and the surface layer 33 can also be determined by a similar procedure.

[0037] The average particle size of the aggregate particles in the surface layer 33 (i.e., the median diameter in the volume-based particle size distribution (D ５０ )) is smaller than the average particle size of the aggregate particles in the intermediate layer 32. The average particle size of the aggregate particles in the intermediate layer 32 is also smaller than the average particle size of the aggregate particles in the base material 31. The average particle size of the aggregate particles in the base material 31 is, for example, 10 μm to 100 μm. The average particle sizes of the aggregate particles in the base material 31, intermediate layer 32, and surface layer 33 can be measured, for example, by laser diffraction.

[0038] As described above, the carbon membrane 12 is formed on the inner surface of each cell 111 (i.e., on the inner surface of the surface layer 33) and covers the inner surface over substantially the entire surface. In this embodiment, the carbon membrane 12 has a substantially cylindrical shape centered on a central axis extending in the longitudinal direction of the cell 111. The carbon membrane 12 is a porous membrane having micropores formed almost entirely of carbon or carbide. The carbon membrane 12 separates a specific substance from a mixture containing multiple substances by using the molecular sieve effect. For example, when a mixture of water and ethanol is introduced into the cell 111 of the carbon membrane composite 1, the water is selectively separated from the mixture by the carbon membrane 12, permeates the support 11, and is discharged from the outer surface of the carbon membrane composite 1.

[0039] The average pore diameter of the carbon membrane 12 is smaller than the average pore diameter of the surface layer 33 of the support 11. The average pore diameter of the carbon membrane 12 is, for example, 0.2 nm or more and 1.0 nm or less. By setting the average pore diameter of the carbon membrane 12 to 0.2 nm or more, it is possible to suppress a decrease in the permeation flux of the substance to be separated. Furthermore, by setting the average pore diameter of the carbon membrane 12 to 1.0 nm or less, it is possible to suppress a decrease in the separation performance (i.e., selectivity) of the carbon membrane 12. The average pore diameter of the carbon membrane 12 can be measured, for example, by gas permeation analysis.

[0040] The thickness of the carbon membrane 12 (i.e., the thickness in the thickness direction) is, for example, 0.1 μm or more and 5.0 μm or less, preferably 0.2 μm or more and 2.5 μm or less, and more preferably 0.3 μm or more and 1.5 μm or less. Increasing the thickness of the carbon membrane 12 improves separation performance. Increasing the thickness of the carbon membrane 12 increases permeation flux.

[0041] The surface roughness of the carbon membrane 12 is, for example, 100 nm or more and 300 nm or less, and preferably 100 nm or more and 200 nm or less. Increasing the surface roughness of the carbon membrane 12 increases the contact area between the carbon membrane 12 and the mixed liquid or the like, thereby increasing the permeation flux of the carbon membrane composite 1. Reducing the surface roughness of the carbon membrane 12 suppresses the occurrence of cracks in the carbon membrane 12 originating from protrusions on the surface of the carbon membrane 12. As a result, the separation performance of the carbon membrane composite 1 is improved.

[0042] Fig. 4 is an enlarged cross-sectional view showing a portion near the carbon membrane 12 of the carbon membrane composite 1. In Fig. 4, a region inside the surface layer 33 of the support 11 where voids 333 between aggregate particles 332 constituting the surface layer 33 are filled with carbon and / or carbide (hereinafter also referred to as "composite layer 13") is marked with parallel hatching that differs from that of the carbon membrane 12. In the composite layer 13, the aggregate particles 332 that are connected by sintering and the voids 333 filled with carbon and / or carbide are marked with the same parallel hatching, and the voids 333 are indicated by thin lines.

[0043] 4, in the region of the surface layer 33 excluding the composite layer 13, the aggregate particles 332 constituting the surface layer 33 are shaded with hatching that differs from that of the carbon membrane 12 and the composite layer 13. In the region of the surface layer 33 excluding the composite layer 13, the voids 333 (i.e., pores) between the aggregate particles 332 of the surface layer 33 are not shaded. In FIGS. 5 and 7, which will be described later, the same hatching as in FIG. 4 is used.

[0044] The composite layer 13 is a region that extends from the surface of the support 11 toward the inside of the support 11 (i.e., radially outward from the inner surface 331 of the surface layer 33). As will be described later, the composite layer 13 is a layer that is formed when the carbon membrane composite 1 is manufactured by performing a carbonization treatment on a precursor solution that will become the carbon membrane 12 in a state where the precursor solution has permeated into a portion of the surface layer 33 near the inner surface 331. In the composite layer 13, carbon and / or carbide (i.e., at least one of carbon and carbide) that is substantially the same as that of the carbon membrane 12 is present on the surfaces of the aggregate particles 332, and carbon and / or carbide is present in a state where it is filled throughout substantially the entire voids 333.

[0045] The thickness of the composite layer 13 (i.e., the thickness in the thickness direction) is not necessarily constant in the longitudinal direction, which is the left-right direction in FIG. 4 , and in the example shown in FIG. 4 , it varies depending on the position in the longitudinal direction. The arithmetic mean of the thickness of the composite layer 13 (hereinafter referred to as the “average thickness t ａｖｅ The average thickness t of the composite layer 13 is, for example, 0.01 μm or more, and preferably 0.05 μm or more. ａｖｅ is, for example, 1.0 μm or less, preferably 0.8 μm or less, more preferably 0.5 μm or less, and further preferably 0.2 μm or less. ｓｄ The standard deviation t of the thickness of the composite layer 13 is, for example, 0.1 μm or more and 1.0 μm or less. ｓｄ It is also preferable that the thickness is 0.5 μm or more.

[0046] Average thickness t of the composite layer 13 ａｖｅ By making the average thickness t of the composite layer 13 1.0 μm or less, the pressure loss that occurs when water or the like permeates through the composite layer 13 is reduced, and the permeation flux of the carbon membrane composite 1 is increased.ａｖｅ By making the thickness 0.01 μm or more, the bond strength between the carbon film 12 and the support 11 is suitably ensured, and peeling of the carbon film 12 from the support 11 is suppressed.

[0047] Standard deviation t of the thickness of the composite layer 13 ｓｄ By setting the standard deviation t of the thickness of the composite layer 13 to 1.0 μm or less, the variation in the thickness of the composite layer 13 is suitably suppressed. Therefore, when water or the like permeates the composite layer 13, the concentration of water or the like in a thin region of the composite layer 13 and an increase in pressure loss are suppressed. As a result, the permeation flux of the carbon membrane composite 1 is increased. In addition, the standard deviation t of the thickness of the composite layer 13 is ｓｄ By setting the standard deviation t to 0.1 μm or more, the thickness variation of the composite layer 13 is allowed to some extent, and the standard deviation t ｓｄ By setting the thickness to 0.5 μm or more, variations in the thickness of the composite layer 13 are further tolerated. Therefore, when manufacturing the carbon membrane composite 1 described later, the thickness of the composite layer 13 can be easily controlled. As a result, the manufacturing of the carbon membrane composite 1 is simplified.

[0048] Average thickness t of the composite layer 13 ａｖｅ , and the standard deviation t of the thickness of the composite layer 13 ｓｄ is determined as follows. First, a polished cross section of a sample of the carbon membrane composite 1 is imaged with an SEM at a magnification of 5000 times or more to obtain a cross-sectional image 91 as shown in FIG. 4 . The cross-sectional image 91 includes the carbon film 12, the boundary between the carbon film 12 and the surface layer 33 (i.e., the inner surface 331 of the surface layer 33), and a region near the inner surface 331 of the surface layer 33. The boundary between the carbon film 12 and the surface layer 33, etc., are determined by the difference in contrast of the cross-sectional image 91.

[0049] Next, as shown in FIG. 5 , the cross-sectional image 91 is divided into a plurality of divided regions 93 by a plurality of straight lines 92 arranged at equal intervals in the longitudinal direction of the carbon membrane composite 1 (i.e., the left-right direction in FIG. 5 ). Each straight line 92 extends parallel to the thickness direction (i.e., the up-down direction in FIG. 5 ) and cuts across the cross-sectional image 91. The plurality of divided regions 93 are arranged in the longitudinal direction. Each divided region 93 is a substantially strip-shaped region extending substantially parallel to the thickness direction (i.e., a substantially rectangular shape elongated in the thickness direction). The widths of the plurality of divided regions 93 in the longitudinal direction are the same. The number of the plurality of divided regions 93 is, for example, 20 or more, and is 20 in this embodiment.

[0050] Next, in each divided region 93, the minimum value of the thickness of the composite layer 13 in the thickness direction is acquired as the thickness of the composite layer 13 in that divided region 93. In other words, in each divided region 93, the distance corresponding to the minimum distance in the thickness direction between the inner surface 331 of the surface layer 33 and the radial outer edge of the composite layer 13 (i.e., the lower edge in FIG. 5 ) is acquired as the thickness of the composite layer 13 in that divided region 93.

[0051] Specifically, first, among the voids 333 that are not filled with carbon and / or carbide in a region below the carbon film 12 in one divided region 93, one void 333 located at the top in FIG. 5 is selected. The selection of the void 333 is performed based on a difference in contrast in the cross-sectional image 91. The upper end of the void 333 in FIG. 5 is determined as a lower end point 94 of the composite layer 13 in the divided region 93. Next, a line passing through the lower end point 94 and parallel to the thickness direction is extended in the vertical direction in FIG. 5, and the point where the line intersects with the inner surface 331 of the surface layer 33 is determined as an upper end point 95 of the composite layer 13 in the divided region 93. The distance in the thickness direction between the upper end point 95 and the lower end point 94 is obtained as the thickness t of the composite layer 13 in the divided region 93. The distance is calculated, for example, in pixel units in the cross-sectional image 91. Note that the distance may be calculated in units of a predetermined pitch, such as 1 μm.

[0052] When the thickness t of the composite layer 13 in each of the divided regions 93 is determined in one polished cross section of the sample of the carbon membrane composite 1, the thickness t of the composite layer 13 in each of the divided regions 93 is also determined in two or more other polished cross sections provided on the sample in the same manner as described above. In this embodiment, the thickness t of the composite layer 13 in each of the 20 divided regions 93 is determined in each of the three polished cross sections. Then, the arithmetic mean of the thicknesses t of the composite layer 13 in a total of 60 divided regions 93 is determined as the average thickness t of the composite layer 13. ａｖｅ In addition, the standard deviation t of the thickness of the composite layer 13 is calculated from the thickness t of the composite layer 13 in the 60 divided regions 93. ｓｄ is required.

[0053] In the carbon membrane composite 1, for the plurality of divided regions 93, the ratio (D50-D10) / (D90-D10) in the distribution of the thickness t of the composite layer 13 is, for example, 0.2 or more and 0.4 or less. In the following description, this ratio (D50-D10) / (D90-D10) is also referred to as the "thickness distribution coefficient." The thickness distribution coefficient is the ratio of the average thickness t of the composite layer 13 to the average thickness t of the composite layer 13. ａｖｅ and standard deviation t ｓｄ The frequency distribution of the thickness t of the composite layer 13 in the 60 divided regions 93 is calculated based on the distribution of the thickness t of the composite layer 13 in the 60 divided regions 93. Fig. 6 is a diagram showing an example of the frequency distribution of the thickness t of the composite layer 13 in the 60 divided regions 93. The horizontal axis in Fig. 6 indicates the thickness t of the composite layer 13 in the divided region 93, the vertical axis on the left indicates the frequency of the thickness t (i.e., the frequency of occurrence), and the vertical axis on the right indicates the cumulative frequency of the thickness t. In Fig. 6, the bar graph indicates the frequency of occurrence, and the line graph indicates the cumulative frequency.

[0054] In the distribution of the thickness t of the composite layer 13 illustrated in FIG. 6 , the thickness distribution coefficient corresponds to L1 / L2 in FIG. 6 . In the composite layer 13, the thickness distribution coefficient decreases as the number of relatively thin regions increases, whereas the thickness distribution coefficient increases as the number of relatively thick regions increases. In the carbon membrane composite 1, by setting the thickness distribution coefficient of the composite layer 13 to 0.4 or less, the number of relatively thin regions increases in the composite layer 13, thereby reducing the pressure drop that occurs when water or the like permeates the composite layer 13 and increasing the permeation flux of the carbon membrane composite 1. Furthermore, by setting the thickness distribution coefficient to 0.2 or more, the number of relatively thin regions in the composite layer 13 is prevented from becoming excessively large. This prevents a decrease in the separation performance of the carbon membrane composite 1. Furthermore, when water or the like permeates the composite layer 13, an increase in pressure drop and a decrease in the permeation flux due to excessive concentration of water or the like in the thin regions of the composite layer 13 are prevented.

[0055] In the following description, as shown by the dashed double-dashed line in FIG. 7, the average thickness t ａｖｅ The region spanning the same distance as the average thickness t is referred to as the "first region 61." In this case, the position in the thickness direction of the inner surface 331 of the surface layer 33 is determined by determining the distance in the thickness direction from the upper end of the image 91 to the inner surface 331 of the surface layer 33 in each divided region 93 shown in FIG. 5 and calculating the arithmetic mean of the distance. The first region 61 is determined when the thickness of the composite layer 13 at each position in the longitudinal direction is equal to or greater than the average thickness t ａｖｅ This corresponds to a virtual composite layer 13 when it is assumed that the thickness of the composite layer 13 is the same as that of the composite layer 13.

[0056] In the following description, the average thickness t of the composite layer 13 in the thickness direction from the surface of the support 11 is ａｖｅ The area extending four times the average thickness t of the composite layer 13 is called the "second area 62." The area obtained by excluding the first area 61 from the second area 62 is called the "composite layer lower area 63." The thickness of the composite layer lower area 63 in the thickness direction is 1 / 2 times the average thickness t of the composite layer 13. ａｖｅ 7, the thickness of the composite layer 13 is not the same at each position in the longitudinal direction, and therefore, the composite layer lower region 63 includes a part of the actual composite layer 13.

[0057] In the carbon membrane composite 1, the porosity of the composite layer lower region 63 is, for example, 5% or more, and preferably 10% or more. The porosity of the composite layer lower region 63 is, for example, 25% or less, and preferably 20% or less. The porosity of the composite layer lower region 63 is calculated by dividing the total area of ​​the voids 333 contained in the composite layer lower region 63 that are not filled with carbon and carbide (i.e., the voids 333 that are not contained in the composite layer 13) by the total area of ​​the composite layer lower region 63. The area of ​​each void 333 was calculated by binarizing the cross-sectional image 91 and then counting the number of pixels contained in each void 333.

[0058] By setting the porosity of the composite layer lower region 63 to 5% or more, variations in the thickness of the composite layer 13 are suitably suppressed. Therefore, when water or the like permeates the composite layer 13, the concentration of water or the like in a thin region of the composite layer 13 and an increase in pressure drop are suppressed. As a result, the permeation flux of the carbon membrane composite 1 is increased. Furthermore, by setting the porosity of the composite layer lower region 63 to 25% or less, variations in the thickness of the composite layer 13 are tolerated to a certain extent. Therefore, during the production of the carbon membrane composite 1 described below, the thickness of the composite layer 13 is easily controlled. As a result, the production of the carbon membrane composite 1 is simplified. Of all the voids 333 in the composite layer lower region 63, the proportion of voids 333 not filled with carbon and / or carbide is preferably 20% to 80%.

[0059] In the carbon membrane composite 1, the ratio of carbon (C) to aluminum (Al) (hereinafter also referred to as "C / Al") in the second region 62 (i.e., the region combining the first region 61 and the composite layer lower region 63) is, for example, 0.05 or more and 0.2 or less.

[0060] The C / Al ratio in the second region 62 can be determined as follows using the graph shown in FIG. 8 . FIG. 8 is a graph showing the results of elemental mapping analysis of the polished cross section 91 (see FIG. 4 ) of the above-described sample using a field emission electron probe microanalyzer (FE-EPMA), where the arithmetic mean of the detected intensity at each depth within the field of view was calculated. The magnification during elemental mapping analysis was 2000 times or more. The horizontal axis in FIG. 8 indicates the measurement position in the thickness direction (i.e., the vertical direction in FIG. 4 ). The vertical axis in FIG. 8 indicates the signal intensity of the characteristic X-rays. The solid line 81 in FIG. 8 indicates the arithmetic mean of the signal intensity of the characteristic X-rays of aluminum (Al), and the solid line 82 indicates the arithmetic mean of the signal intensity of the characteristic X-rays of carbon (C). In the example shown in FIG. 8 , the solid line 81 includes an increasing region that gradually increases as the measurement position increases from 0, and a constant value region that extends linearly to the right at a substantially constant value from the right end of the increasing region. The increase region of the solid line 81 includes a first increase region in which the slope of the solid line 81 gradually increases as the measurement position increases from 0, a second increase region in which the slope of the solid line 81 extends in an approximately straight line diagonally upward to the right from the right end of the first increase region with an approximately constant slope, and a third increase region in which the slope of the solid line 81 gradually decreases as the measurement position increases from the right end of the second increase region.

[0061] When determining the C / Al ratio in the second region 62, first, an inflection point 83 of the solid line 81 is determined, and the measurement position corresponding to the inflection point 83 is defined as the surface of the support 11 (i.e., the inner surface 331 of the surface layer 33). The inflection point 83 is determined by determining the intersection of an approximation line corresponding to the second increasing region and an approximation line corresponding to the constant value region (i.e., a line parallel to the horizontal axis), and is determined as a point on the solid line 81 at the measurement position of the intersection. The region to the right of the measurement position corresponding to the inflection point 83 in FIG. 8 corresponds to the interior of the support 11, and the right end of the horizontal axis corresponds to the radial outer end of the second region 62. Then, in FIG. 8 , the C / Al ratio in the second region 62 is determined by dividing the arithmetic mean of the signal intensity of the characteristic X-rays of carbon within the support 11 by the arithmetic mean of the signal intensity of the characteristic X-rays of aluminum within the support 11.

[0062] By setting the C / Al in the second region 62 to 0.05 or more, the density of carbon in the second region 62 is increased, improving the separation performance of the carbon membrane composite 1. By setting the C / Al in the second region 62 to 0.2 or less, the carbon content in the second region 62 is prevented from becoming excessively high. This reduces the pressure loss that occurs when water or the like permeates through the second region 62, and increases the permeation flux of the carbon membrane composite 1.

[0063] Next, the production of the carbon membrane composite 1 will be described with reference to Fig. 9 and Fig. 10A to Fig. 10C. Fig. 9 is a diagram showing an example of the flow of producing the carbon membrane composite 1 by the first production method. Fig. 10A to Fig. 10C are diagrams schematically showing a partial cross section of the carbon membrane composite 1 in the process of production.

[0064] When the carbon membrane composite 1 is produced by the first production method, first, the support 11 is prepared (step S11). In step S11, for example, the raw material of the substrate 31 is molded by extrusion molding or the like and then fired to form the substrate 31 having a monolith structure. Next, the intermediate layer 32 and the surface layer 33 are formed on the inner surfaces of the through holes of the substrate 31 by filtration membrane formation or the like and then fired to form and prepare the support 11.

[0065] Next, a predetermined non-solvent is applied to the support 11 and allowed to soak in. As a result, as shown in Fig. 10A, a non-solvent layer 71 in which the non-solvent exists between the aggregate particles 332 is formed inside the support 11 (step S12). In Fig. 10A, the non-solvent layer 71 is indicated by hatching (the same applies to Figs. 10B to 10C). In the example shown in Fig. 10A, the non-solvent has soaked into almost the entire support 11, and the non-solvent layer 71 is provided over almost the entire surface layer 33, almost the entire intermediate layer 32, and almost the entire base material 31 (not shown).

[0066] The non-solvent is a liquid in which the precursor to be formed into the carbon film 12 has low solubility, and is a different type of liquid from the solvent of the precursor solution described below. The non-solvent is a liquid in which the solubility of the precursor at 25°C (i.e., the upper limit of the mass of the precursor that dissolves in 100 g of the non-solvent) is 0.1 g or less. In this embodiment, water, which is easily available, is used as the non-solvent.

[0067] The application of the non-solvent in step S12 may be performed, for example, by storing the non-solvent in an open-top container and immersing substantially the entire support 11 in the non-solvent in the container in an atmospheric environment (hereinafter also referred to as the "immersion method"). Alternatively, the application of the non-solvent in step S12 may be performed by placing the container in the internal space of a degassing device with substantially the entire support 11 immersed in the non-solvent in the container, and evacuating the internal space of the degassing device using a vacuum pump or the like (hereinafter also referred to as the "immersion degassing method"). Alternatively, the application of the non-solvent in step S12 may be performed in an atmospheric environment by placing the support 11 so that the longitudinal direction of the support 11 is substantially parallel to the direction of gravity, and allowing the non-solvent to flow in from the upper opening of each cell 111 (hereinafter also referred to as the "flow-down method").

[0068] After step S12 is completed, as shown in FIG. 10B , a portion of the non-solvent layer 71 provided inside the support 11 is removed (step S13). The partial removal of the non-solvent layer 71 in step S13 is performed, for example, by a drying process such as blowing air onto the support 11. This drying process vaporizes and removes a portion of the non-solvent constituting the non-solvent layer 71 from the inner surface of each cell 111 (i.e., the inner surface 331 of the surface layer 33) of the cell 111. This removes the radially inner portion of the non-solvent layer 71 (i.e., the portion near the inner surface 331 of the surface layer 33), and the radially inner end of the non-solvent layer 71 moves radially outward from the inner surface of the cell 111. After step S13 is completed, the radially inner end of the non-solvent layer 71 is located radially inward of the interface between the surface layer 33 and the intermediate layer 32 and near the inner surface 331 of the surface layer 33.

[0069] The drying process in step S13 is performed, for example, at room temperature (25° C.) for a predetermined time. In step S13, the amount of non-solvent layer 71 removed from the inner surface of each cell 111 is changed by changing the drying time (i.e., drying time). In other words, the distance between the inner surface 331 of each cell 111 and the radially inner end of the non-solvent layer 71 in each cell 111 is adjusted by changing the drying time. Specifically, by extending the drying time, the distance between the inner surface of each cell 111 and the radially inner end of the non-solvent layer 71 increases, and the region in the support 11 where no non-solvent is present expands radially outward as viewed from the inner surface of each cell 111.

[0070] When step S13 is completed, the periphery of the outer surface of the support 11 (i.e., the space radially outward from the outer surface of the support 11) is pressurized. Specifically, for example, a substantially cylindrical casing made of stainless steel or the like covers substantially the entire outer surface of the support 11. Because the inner diameter of the casing is larger than the outer diameter of the support 11, a substantially cylindrical space (hereinafter also referred to as the "outer peripheral space") exists between the inner surface of the casing and the outer surface of the support 11 around the entire circumference. At both longitudinal ends of the support 11, O-rings or the like are provided between the inner surface of the casing and the outer surface of the support 11, and the outer peripheral space is airtightly sealed. The outer peripheral space is pressurized by a pressure pump or the like connected to the casing. Specifically, the outer peripheral space is pressurized by supplying pressurized gas (e.g., air) into the casing. As a result, pressure increases in the surface layer 33 and intermediate layer 32 of the support 11 from the inner surface 331 of the surface layer 33 toward the radially outward direction. The pressure in the outer space is, for example, 10 kPaG to 500 kPaG.

[0071] Then, with the outer peripheral space pressurized, a precursor solution is applied to the support 11 from the inner surface of the cell 111 (i.e., the inner surface 331 of the surface layer 33). The precursor solution is a liquid containing a precursor that will become carbon in the carbon film 12. As shown in FIG. 10C , the precursor solution permeates the support 11 from the inner surface 331 of the surface layer 33 and spreads to the region where the non-solvent layer 71 was removed in step S13 (i.e., the region near the inner surface 331 of the surface layer 33), forming a precursor layer 72. The precursor solution also adheres to the inner surface 331 of the surface layer 33. As a result, a precursor film 73, which is a thin film containing the precursor, is formed on the inner surface 331 of the surface layer 33 (i.e., the surface of the support 11) within which the precursor layer 72 has been formed (step S14). In FIG. 10C , the precursor layer 72 and the precursor film 73 are indicated by the same diagonal lines.

[0072] The precursor may be, for example, a thermosetting resin such as a phenolic resin, a melamine resin, a urea resin, a furan resin, a polyimide resin, or an epoxy resin; a thermoplastic resin such as polyethylene, polyphenylene oxide, or polyetherimide; a cellulose-based resin; or an intrinsically microporous polymer (PIM). Alternatively, the precursor may be a precursor material of these resins. In this embodiment, a phenolic resin is used as the precursor.

[0073] The precursor solution is a solution in which a precursor is mixed or dissolved in a solvent. The solvent is, for example, an organic solvent such as methanol, ethanol, acetone, tetrahydrofuran, NMP (N-methyl-2-pyrrolidone), or toluene, or water. In this embodiment, ethanol is used as the solvent. Note that the precursor and solvent contained in the precursor solution are not limited to the above examples and may be variously changed.

[0074] The precursor solution is applied by, for example, a flow-down method, an immersion method, dip coating, spin coating, drip coating, spray coating, or filtration coating. In this embodiment, the precursor solution is applied by a flow-down method. Specifically, the support 11 is placed so that the longitudinal direction of the support 11 is approximately parallel to the direction of gravity, and the precursor solution is poured into the upper openings of each cell 111, thereby applying the precursor solution to the support 11.

[0075] In the example shown in FIG. 10C , the radially outer end of the precursor layer 72 contacts the radially inner end of the non-solvent layer 71. As described above, the non-solvent constituting the non-solvent layer 71 differs from the solvent of the precursor solution constituting the precursor layer 72, and the solubility of the precursor in the non-solvent is low. Therefore, the precursor in the precursor solution rapidly aggregates at the interface between the non-solvent layer 71 and the precursor layer 72, suppressing diffusion of the precursor solution into the non-solvent layer 71. Furthermore, in step S14, the precursor layer 72 and the precursor film 73 are formed in a state in which pressure is applied from the outer peripheral space side of the support 11 (i.e., the side opposite the inner surface 331 of the surface layer 33 of the support 11). This suppresses radially outward displacement of the radially outer end of the precursor layer 72. As a result, unintended increases in the radial thickness of the precursor layer 72 are suppressed. The radial thickness of the precursor layer 72 can be adjusted by adjusting the degree of pressure applied to the outer peripheral space.

[0076] The support 11 provided with the precursor layer 72 and precursor film 73 is dried while maintaining the pressurized outer peripheral space (step S15). The support 11 is dried, for example, by blowing air at room temperature (25°C) for 20 minutes. Since the non-solvent layer 71 inside the surface layer 33 and the intermediate layer 32 is covered by the precursor layer 72 and precursor film 73 from the radially inner side, it is not significantly removed by the drying process. Therefore, the position of the radially inner end of the non-solvent layer 71 does not substantially move, and the position of the radially outer end of the precursor layer 72 does not substantially move, either. Furthermore, in step S15, the precursor layer 72 and precursor film 73 are dried while applying pressure from the outer peripheral space side of the support 11 (i.e., the side of the support 11 opposite the inner surface 331 of the surface layer 33). This prevents the radially outer end of the precursor layer 72 from being displaced radially outward. As a result, an unintended increase in the radial thickness of the precursor layer 72 is prevented.

[0077] The support 11 after the above-described drying is subjected to a heat treatment to remove the non-solvent layer inside the support 11 and to harden the precursor film 73 on the surface of the support 11 and the precursor layer 72 inside the support 11. The heat treatment is performed, for example, by heating the support 11 to 90°C to 500°C using a dryer and holding the temperature for 0.5 hours to 60 hours.

[0078] When forming the precursor film 73, if the film thickness of the precursor film 73 does not reach the desired thickness by applying the precursor solution and performing the above-mentioned heating treatment only once, the precursor film 73 shown in Figure 7C may be formed by repeating the above-mentioned application of the precursor solution and heating treatment multiple times.

[0079] When the curing of the precursor layer 72 and the precursor film 73 is completed, the support body 11 provided with the precursor layer 72 and the precursor film 73 is subjected to a carbonization treatment. As a result, a carbon film 12 is formed from the precursor film 73 on the surface of the support body 11 (i.e., the inner surface 331 of the cell 111). Furthermore, a composite layer 13 is formed from the precursor layer 72 in a portion of the interior of the support body 11 near the surface. As a result, a carbon membrane composite 1 including the carbon film 12 and the composite layer 13 is obtained (step S16).

[0080] In the carbonization treatment in step S16, for example, the support 11 provided with the precursor layer 72 and the precursor film 73 is placed in an electric furnace and heated to 400°C to 1200°C. By setting the temperature during the carbonization treatment to 400°C or higher, the precursor film 73 is suitably carbonized, pores are formed, and the separation performance of the carbon film 12 is suitably exhibited. By setting the temperature during the carbonization treatment to 1200°C or lower, the carbon film 12 is prevented from becoming excessively dense, and the separation performance of the carbon film 12 is suitably exhibited. It is more preferable that the temperature during the carbonization treatment be 600°C to 900°C.

[0081] The carbonization treatment is preferably carried out in a non-oxidizing atmosphere. The non-oxidizing atmosphere means an atmosphere in which the precursor is not oxidized when carbonized within the temperature range. Specifically, the non-oxidizing atmosphere is an inert gas atmosphere such as nitrogen or argon, or a vacuum atmosphere.

[0082] Next, separation of a mixed substance using the carbon membrane composite 1 will be described with reference to Fig. 11. Fig. 11 is a cross-sectional view showing a separation device 2. In Fig. 11, the cross section of the carbon membrane composite 1 is shown conceptually in a simplified form in order to facilitate understanding of the drawing.

[0083] In the separation device 2, a mixed substance containing multiple types of fluids (i.e., gas or liquid) is supplied to the carbon membrane composite 1, and highly permeable substances in the mixed substance are separated from the mixed substance by permeating the carbon membrane composite 1. Separation in the separation device 2 may be performed, for example, for the purpose of extracting highly permeable substances (hereinafter also referred to as "highly permeable substances") from the mixed substance, or for the purpose of concentrating lowly permeable substances (hereinafter also referred to as "lowly permeable substances").

[0084] The mixed substance (i.e., mixed fluid) may be a mixed gas containing multiple types of gases, a mixed liquid containing multiple types of liquids, or a gas-liquid two-phase fluid containing both gas and liquid.

[0085] The mixture may include, for example, hydrogen (H ２ ), helium (He), nitrogen (N ２ ), oxygen (O ２), water (H ２ O), carbon monoxide (CO), carbon dioxide (CO ２ ), nitrogen oxides, ammonia (NH ３ ), sulfur oxides, hydrogen sulfide (H ２ S), sulfur fluoride, mercury (Hg), arsine (AsH ３ ), hydrogen cyanide (HCN), carbonyl sulfide (COS), C1-C8 hydrocarbons, organic acids, alcohols, mercaptans, esters, ethers, ketones, and aldehydes. ２ , N.H. ３ and H ２ The mixed substance and the highly permeable substance may be one or more of the substances listed above.

[0086] Nitrogen oxides are compounds of nitrogen and oxygen. Examples of the nitrogen oxides include nitric oxide (NO), nitrogen dioxide (NO ２ ), nitrous oxide (also called dinitrogen monoxide) (N ２ O), dinitrogen trioxide (N ２ O ３ ), dinitrogen tetroxide (N ２ O ４ ), dinitrogen pentoxide (N ２ O ５ ) etc. Ｘ It is a substance called NOX.

[0087] Sulfur oxides are compounds of sulfur and oxygen. Examples of the sulfur oxides include sulfur dioxide (SO ２ ), sulfur trioxide (SO ３ ) and other SOs Ｘ It is a substance called SOXI.

[0088] Sulfur fluoride is a compound of fluorine and sulfur. The above-mentioned sulfur fluoride is, for example, disulfur difluoride (FS-SF, S=SF ２ ), sulfur difluoride (SF ２ ), sulfur tetrafluoride (SF ４ ), sulfur hexafluoride (SF ６ ) or disulfur decafluoride (S ２ F １０ ) etc.

[0089] C1-C8 hydrocarbons are hydrocarbons with one or more carbon atoms and eight or less. C3-C8 hydrocarbons may be straight-chain compounds, branched-chain compounds, or cyclic compounds. C2-C8 hydrocarbons may be either saturated hydrocarbons (i.e., those without double and triple bonds in the molecule) or unsaturated hydrocarbons (i.e., those with double and / or triple bonds in the molecule). C1-C4 hydrocarbons include, for example, methane (CH ４ ), ethane (C ２ H ６ ), ethylene (C ２ H ４ ), propane (C ３ H ８ ), propylene (C ３ H ６ ), normal butane (CH ３ (CH ２ ) ２ CH ３ ), isobutane (CH(CH ３ ) ３ ), 1-butene (CH ２ =CHCH ２ CH ３ ), 2-butene (CH ３ CH=CHCH ３ ) or isobutene (CH ２ =C(CH ３ ) ２ )

[0090] The organic acid is a carboxylic acid or a sulfonic acid. The carboxylic acid is, for example, formic acid (CH ２ O ２ ), acetic acid (C ２ H ４ O ２ ), oxalic acid (C ２ H ２ O ４ ), acrylic acid (C ３ H ４ O ２ ) or benzoic acid (C ６ H ５ COOH), etc. Sulfonic acids include, for example, ethanesulfonic acid (C ２ H ６ O ３S) and the like. The organic acid may be a chain compound or a cyclic compound.

[0091] The alcohols mentioned above include, for example, methanol (CH ３ OH), ethanol (C ２ H ５ OH), isopropanol (2-propanol) (CH ３ CH(OH)CH ３ ), ethylene glycol (CH ２ (OH)CH ２ (OH)), butanol (C ４ H ９ OH) or phenol (C ６ H ５ OH) and the like.

[0092] Mercaptans are organic compounds with hydrogenated sulfur (SH) at the end, and are also called thiols or thioalcohols. Examples of the above-mentioned mercaptans include methyl mercaptan (CH ３ SH), ethyl mercaptan (C ２ H ５ SH) or 1-propanethiol (C ３ H ７ SH), etc.

[0093] The esters mentioned above are, for example, formates or acetates.

[0094] The above-mentioned ethers include, for example, dimethyl ether ((CH ３ ) ２ O), methyl ethyl ether (C ２ H ５ OCH ３ ), diethyl ether ((C ２ H ５ ) ２ O), tetrahydrofuran (C ４ H ８ O) or propylene glycol methyl ether (CH ３ OCH ２ CH(OH)CH ３ ) etc.

[0095] The above-mentioned ketones can be, for example, acetone ((CH ３ )２ CO), methyl ethyl ketone (C ２ H ５ COCH ３ ), diethyl ketone ((C ２ H ５ ) ２ CO) or N-methyl-2-pyrrolidone (C ５ H ８ NOCH ３ ) etc.

[0096] The aldehydes mentioned above include, for example, acetaldehyde (CH ３ CHO), propionaldehyde (C ２ H ５ CHO) or butanal (butyraldehyde) (C ３ H ７ CHO) etc.

[0097] As shown in Fig. 11 , the separation device 2 includes a carbon membrane composite 1, a sealing unit 21, a housing 22, and two sealing members 23. The carbon membrane composite 1, the sealing unit 21, and the sealing members 23 are housed in the housing 22. In Fig. 11 , the support 11 and the carbon membrane 12 of the carbon membrane composite 1 are indicated by parallel diagonal lines. The internal space of the housing 22 is a sealed space isolated from the space surrounding the housing 22. A supply unit 26, a first recovery unit 27, and a second recovery unit 28 are connected to the housing 22.

[0098] As described above, the sealing portion 21 is attached to both ends of the support 11 in the longitudinal direction (i.e., the left-right direction in FIG. 11 ), and is a member that covers and seals both end faces of the support 11 in the longitudinal direction and parts of the outer surfaces near these end faces. In this embodiment, the sealing portion 21 is a glass seal with a thickness of 10 μm to 50 μm. The material and shape of the sealing portion 21 may be changed as appropriate. Note that the sealing portion 21 has multiple openings that overlap with the multiple cells 111 of the support 11, and therefore both ends of each cell 111 in the longitudinal direction are not covered by the sealing portion 21. Therefore, fluid can flow in and out of the cells 111 from these ends.

[0099] The housing 22 is a substantially cylindrical tubular member. The housing 22 is made of, for example, stainless steel or carbon steel. The longitudinal direction of the housing 22 is substantially parallel to the longitudinal direction of the carbon membrane composite 1. A supply port 221 is provided at one longitudinal end of the housing 22 (i.e., the left end in FIG. 11 ), and a first discharge port 222 is provided at the other end. A supply unit 26 is connected to the supply port 221. A first recovery unit 27 is connected to the first discharge port 222. A second discharge port 223 is provided on a side surface of the housing 22. A second recovery unit 28 is connected to the second discharge port 223. The shape and material of the housing 22 may be modified in various ways.

[0100] Two seal members 23 are disposed between the outer surface of the carbon membrane composite 1 and the inner surface of the housing 22 near both longitudinal ends of the carbon membrane composite 1. Each seal member 23 is a substantially annular member made of a material impermeable to gases and liquids. The seal members 23 are, for example, O-rings or packings made of flexible resin. The seal members 23 are in close contact with the outer surface of the carbon membrane composite 1 and the inner surface of the housing 22 around the entire circumference of the carbon membrane composite 1. In the example shown in FIG. 11 , the seal member 23 is in close contact with the outer surface of the sealing portion 21 and indirectly with the outer surface of the carbon membrane composite 1 via the sealing portion 21. The seal member 23 may also be in direct contact with the outer surface of the carbon membrane composite 1. A seal is formed between each seal member 23 and the outer surface of the carbon membrane composite 1 or the sealing portion 21, and between each seal member 23 and the inner surface of the housing 22, substantially preventing the passage of gases and liquids. The material of the seal member 23 may be carbon, metal, or other inorganic material other than resin.

[0101] Supply unit 26 supplies the mixed substance to the internal space of housing 22 via supply port 221. Supply unit 26 includes, for example, a pressure-feeding mechanism such as a blower or pump that pressure-feeds the mixed substance toward housing 22. The pressure-feeding mechanism includes, for example, a temperature adjustment unit and a pressure adjustment unit that respectively adjust the temperature and pressure of the mixed substance supplied to housing 22. First recovery unit 27 and second recovery unit 28 include, for example, a storage container that stores the substance discharged from housing 22, or a blower or pump that transports the substance.

[0102] When separating a mixed substance, first, the carbon membrane composite 1 is prepared. Specifically, the carbon membrane composite 1 is attached inside the housing 22. Then, a mixed substance containing multiple types of substances with different permeabilities to the carbon membrane 12 is supplied into the housing 22 by the supply unit 26, as indicated by the arrow 251. For example, the mixed substance is a mixed liquid in which multiple types of liquids are mixed. The main components of the mixed liquid are, for example, water and ethanol. The mixed liquid may contain liquids other than water and ethanol.

[0103] The mixed substance supplied from the supply unit 26 to the housing 22 is introduced into each cell 111 of the support 11 from the left end of the carbon membrane composite 1 in the drawing. A highly permeable substance in the mixed substance, which is a substance with high permeability, permeates through the carbon membrane 12 provided on the inner surface of each cell 111 and the support 11, and is discharged from the outer surface of the support 11. In this way, the highly permeable substance (e.g., water) is separated from a less permeable substance in the mixed substance (e.g., ethanol).

[0104] The substance discharged from the outer surface of the support 11 (hereinafter referred to as the "permeated substance") is led to the second recovery section 28 via the second discharge port 223 as indicated by the arrow 253, and is recovered by the second recovery section 28. The permeated substance may include, in addition to the above-mentioned highly permeable substance, a low-permeable substance that has permeated the carbon membrane 12.

[0105] Furthermore, of the mixed substance, substances other than those that have permeated the carbon membrane 12 and the support 11 (hereinafter referred to as "non-permeated substances") pass through each cell 111 of the support 11 from left to right in the figure and are recovered by the first recovery unit 27 via the first discharge port 222, as indicated by arrow 252. The non-permeated substances may include, in addition to the low-permeable substances described above, highly permeable substances that have not permeated the carbon membrane 12. The non-permeated substances recovered by the first recovery unit 27 may be circulated to the supply unit 26, for example, and supplied again into the housing 22.

[0106] Next, with reference to Tables 1 to 3, Examples 1 to 8 and Comparative Examples 1 to 5 of the carbon membrane composite 1 will be described.

[0107]

[0108]

[0109]

[0110] The carbon membrane composites 1 of Examples 1 to 8 were manufactured by the manufacturing method shown in steps S11 to S16 described above. Examples 1 to 8 differed from one another in the conditions for applying the non-solvent in step S12 and / or the conditions for partially removing the non-solvent layer 71 in step S13. In Examples 1 to 8, the pressure in the outer peripheral space pressurized in steps S14 and S15 was 100 kPaG. The precursor layer 72 and the precursor film 73 were cured by heating the support 11 to 200°C in an air atmosphere and maintaining that temperature for 5 hours. The carbonization treatment in step S16 was performed by heating the support 11 to 850°C in a vacuum atmosphere and maintaining that temperature for 5 hours. The carbon membrane composites of Comparative Examples 1 to 5 were manufactured in substantially the same manner as Examples 1 to 8. However, in Comparative Examples 1, 3, and 4, the application of the non-solvent in step S12 and the partial removal of the non-solvent layer 71 in step S13 were not performed. Furthermore, in Comparative Examples 1 and 3 to 5, the pressure in the outer peripheral space was not applied in steps S14 and S15. That is, the pressure in the outer peripheral space in steps S14 and S15 in Comparative Examples 1 and 3 to 5 was 0 kPaG.

[0111] In Examples 1 to 8 and Comparative Examples 1 to 5, the average pore diameter on the surface of the support 11, the thickness of the carbon film 12, the average thickness and standard deviation of the composite layer 13, the porosity of the composite layer lower region 63, the thickness distribution coefficient of the composite layer 13, and the C / Al ratio in the second region 62 were obtained by the methods described above. The thickness of the carbon film 12 was measured using an SEM "S-3400N" manufactured by Hitachi High-Tech Corporation at an acceleration voltage of 15 kV, a field of view of 36 μm × 50 μm, and a magnification of 5000 times. The C / Al ratio in the second region 62 was obtained using an FE-EPMA "JXA-8500F" manufactured by JEOL Ltd. at an acceleration voltage of 5 kV, a field of view of 38 μm × 50 μm, and a magnification of 2000 times.

[0112] The permeation flux and separation performance in Table 3 were measured using the separation device 2 (see FIG. 11 ). Specifically, a mixed liquid of water and ethanol was first supplied to the carbon membrane composite 1 from the supply section 26 of the separation device 2, and the liquid separated by pervaporation (PV) (i.e., the liquid that permeated the carbon membrane 12 and the support 11; hereinafter, also referred to as the "permeated liquid") was recovered in the second recovery section 28. The mass ratio of water to ethanol in the mixed liquid was 50:50, and the temperature of the mixed liquid supplied from the supply section 26 was 50°C. The permeation-side pressure, which is the pressure on the second recovery section 28 side, was 6.67 kPa (i.e., 50 Torr). Next, the mass and density of the permeated liquid recovered in the second recovery section 28 were measured, and the mass ratio of water to ethanol in the permeated liquid was calculated. The density of the permeated liquid was measured using a density specific gravity meter. Then, from the mass of the permeated liquid and the mass ratio of water to ethanol in the permeated liquid, the water permeation flux (kg / m ２ h) and the separation factor of water relative to ethanol were determined.

[0113] In Table 3, the "excellent" in the permeation flux column indicates that the water permeation flux is 1.2 kg / m ２ h or more, and "good" indicates a water permeation flux of 0.4 kg / m ２ h or more and 1.2 kg / m ２ In addition, "Fail" in the permeation flux column indicates that the water permeation flux is less than 0.4 kg / m ２In the column for separation performance in Table 3, "excellent" indicates that the separation factor is 80 or more, and "good" indicates that the separation factor is 60 or more and less than 80.

[0114] In Example 1, the application of the non-solvent in step S12 was performed by the immersion degassing method. In step S12, water was used as the non-solvent, and the internal space of the degassing device was evacuated and degassed for 20 minutes. Furthermore, in step S13, partial removal of the non-solvent layer 71 was performed by a drying treatment using air blowing. In step S13, the drying treatment time (i.e., drying time) was set to 15 seconds. The average pore diameter on the surface of the support 11 was 0.07 μm.

[0115] In Example 1, the precursor concentration in the precursor solution applied to the support 11 in step S14 was 7.5 mass %. The precursor was a phenolic resin (Bellpearl (registered trademark) S899, manufactured by Air Water Inc.), and the solvent for the precursor solution was ethanol (manufactured by Amakasu Chemical Industry Co., Ltd.). The precursor solution was produced by weighing and mixing the precursor and solvent so that the precursor concentration was the above-mentioned concentration, stirring the mixture at room temperature with a magnetic stirrer for four days, and then filtering the mixture using a 20 μm sieve. In Example 1, the outer peripheral space was pressurized to 100 kPaG in steps S14 and S15.

[0116] In Example 1, the thickness of the carbon film 12 was 1.67 μm. The average thickness t ａｖｅ is 0.52 μm, and the standard deviation of the thickness of the composite layer 13 is t ｓｄ The thickness distribution coefficient of the composite layer 13 (i.e., (D50-D10) / (D90-D10)) was 0.22. The C / Al ratio in the second region 62 was 0.11. In Example 1, the permeation flux was "excellent" and the separation performance was also "excellent".

[0117] Example 2 is similar to Example 1 except for the method of applying the non-solvent in step S12. In Example 2, the non-solvent (i.e., water) was applied to the support 11 by immersion.

[0118] In Example 2, the thickness of the carbon film 12 was 1.70 μm. The average thickness t ａｖｅ is 0.78 μm, and the standard deviation of the thickness of the composite layer 13 is t ｓｄ The thickness distribution coefficient of the composite layer 13 was 0.39. The C / Al ratio in the second region 62 was 0.18. In Example 2, the permeation flux was "good" and the separation performance was "excellent."

[0119] Example 3 is similar to Example 1 except for the method of applying the non-solvent in step S12. In Example 3, the non-solvent (i.e., water) was applied to the support 11 by a dripping method. The amount of the non-solvent poured into each cell 111 was 200 mL (milliliters).

[0120] In Example 3, the thickness of the carbon film 12 was 1.72 μm. The average thickness t ａｖｅ is 0.83 μm, and the standard deviation of the thickness of the composite layer 13 is t ｓｄ The thickness distribution coefficient of the composite layer 13 was 0.33. The C / Al ratio in the second region 62 was 0.16. In Example 3, the permeation flux was "good" and the separation performance was "excellent."

[0121] Example 4 is the same as Example 3, except for the amount of non-solvent poured into the cells 111 in step S12. In Example 4, the amount of non-solvent poured into each cell 111 by the dripping method is 100 mL (milliliters).

[0122] In Example 4, the thickness of the carbon film 12 was 1.70 μm. The average thickness t ａｖｅ is 0.85 μm, and the standard deviation of the thickness of the composite layer 13 is t ｓｄ The thickness distribution coefficient of the composite layer 13 was 0.32. The C / Al ratio in the second region 62 was 0.17. In Example 4, the permeation flux was "good" and the separation performance was "excellent."

[0123] Example 5 is the same as Example 1 except for the drying time in step S13, which was 10 seconds.

[0124] In Example 5, the thickness of the carbon film 12 was 1.71 μm. The average thickness t ａｖｅ is 0.07 μm, and the standard deviation of the thickness of the composite layer 13 is t ｓｄ The thickness distribution coefficient of the composite layer 13 was 0.20. The C / Al ratio in the second region 62 was 0.09. In Example 5, the permeation flux was "excellent" and the separation performance was also "excellent".

[0125] Example 6 is the same as Example 1 except for the degassing time in step S12, which was 30 minutes.

[0126] In Example 6, the thickness of the carbon film 12 was 1.67 μm. The average thickness t ａｖｅ is 0.50 μm, and the standard deviation of the thickness of the composite layer 13 is t ｓｄ The thickness distribution coefficient of the composite layer 13 was 0.18. The C / Al ratio in the second region 62 was 0.10. In Example 6, the permeation flux was "excellent" and the separation performance was also "excellent".

[0127] Example 7 is the same as Example 1 except for the drying time in step S13, which was 5 seconds.

[0128] In Example 7, the thickness of the carbon film 12 was 1.70 μm. The average thickness t ａｖｅ is 0.05 μm, and the standard deviation of the thickness of the composite layer 13 is t ｓｄ The thickness distribution coefficient of the composite layer 13 was 0.21. The C / Al ratio in the second region 62 was 0.04. In Example 7, the permeation flux was "excellent" and the separation performance was "good".

[0129] Example 8 is the same as Example 4, except for the drying time in step S13, which was 20 seconds.

[0130] In Example 8, the thickness of the carbon film 12 was 1.70 μm. The average thickness t ａｖｅ is 0.83 μm, and the standard deviation of the thickness of the composite layer 13 is t ｓｄ The thickness distribution coefficient of the composite layer 13 was 0.34. The C / Al ratio in the second region 62 was 0.25. In Example 8, the permeation flux was "good" and the separation performance was "excellent."

[0131] Comparative Example 1 is similar to Example 1 except that the application of non-solvent in step S12 and the partial removal of non-solvent layer 71 in step S13 are omitted, and the pressure in the outer space in steps S14 and S15 is 0 kPaG.

[0132] In Comparative Example 1, the thickness of the carbon film 12 was 1.73 μm. The average thickness t ａｖｅ The standard deviation of the thickness of the composite layer 13 is t ｓｄ The particle size was also large at 1.2 μm. The porosity of the composite layer lower region 63 was small at 2.3%. The thickness distribution coefficient of the composite layer 13 was 0.56. The C / Al ratio in the second region 62 was 0.30. In Comparative Example 1, the permeation flux was "unacceptable" and the separation performance was "excellent."

[0133] Comparative Example 2 is the same as Example 1 except that the degassing time in step S12 and the drying time in step S13 are different. In Comparative Example 2, the degassing time was 60 minutes, and the drying time was 2 seconds.

[0134] In Comparative Example 2, the thickness of the carbon film 12 was 1.69 μm. In Comparative Example 2, voids that were continuous with the voids 333 of the support 11 were generated inside the carbon film 12, and therefore the average thickness t ａｖｅ , and the standard deviation of the thickness t ｓｄIn Comparative Example 2, the permeation flux and separation performance could not be measured because the carbon membrane 12 peeled off from the support 11 in the separation device 2. In Comparative Example 2, the porosity of the composite layer lower region 63 and the thickness distribution coefficient of the composite layer 13 could not be measured. In addition, the porosity of the composite layer lower region 63 and the thickness distribution coefficient of the composite layer 13 could not be measured. The C / Al in the second region 62 was 0.01. In Comparative Example 2, the permeation flux and separation performance could not be measured because the carbon membrane 12 peeled off from the support 11 in the separation device 2.

[0135] Comparative Example 3 is similar to Comparative Example 1 except that, similarly to the above-mentioned International Publication No. 2008 / 010452, a dense surface deposition layer was provided on the surface of the support 11 and the pressure in the outer peripheral space in steps S14 and S15 was 0 kPaG. In Comparative Example 3, the average pore diameter on the surface of the support 11 was reduced to 0.01 μm, thereby suppressing the penetration of the precursor solution of the carbon film 12 into the support 11.

[0136] In Comparative Example 3, the thickness of the carbon film 12 was 1.74 μm. The average thickness t ａｖｅ The standard deviation t of the thickness of the composite layer 13 was 0.85 μm. ｓｄ was 1.2 μm, which was large as in Comparative Example 1. The porosity of the composite layer lower region 63 was 2.5%, which was small as in Comparative Example 1. The thickness distribution coefficient of the composite layer 13 was 0.52. The C / Al in the second region 62 was 0.17. In Comparative Example 3, the permeation flux was "unacceptable" and the separation performance was "excellent."

[0137] Comparative Example 4 is similar to Comparative Example 1, except that, similarly to the above-mentioned International Publication No. 2013 / 042262, a highly viscous precursor solution was applied to the support 11 in step S14, and the pressure in the outer peripheral space in steps S14 and S15 was 0 kPaG. In Comparative Example 4, the use of a highly viscous precursor solution prevented the precursor solution of the carbon film 12 from permeating into the support 11. In Comparative Example 4, the average pore diameter on the surface of the support 11 was set to 0.1 μm.

[0138] In Comparative Example 4, the thickness of the carbon film 12 was 1.73 μm. The average thickness t ａｖｅ is as large as 2.0 μm, and the standard deviation t ｓｄThe porosity of the composite layer lower region 63 was 2.3%, which was small as in Comparative Example 1. The thickness distribution coefficient of the composite layer 13 was 0.51. The C / Al ratio in the second region 62 was 0.22. In Comparative Example 4, the permeation flux was "unacceptable" and the separation performance was "excellent."

[0139] Comparative Example 5 is similar to Example 1 except that the pressure in the outer peripheral space in steps S14 and S15 is 0 kPaG.

[0140] In Comparative Example 5, the thickness of the carbon film 12 was 1.4 μm. The average thickness t ａｖｅ is large, 1.5 μm, and the standard deviation t ｓｄ The particle size was also large at 1.2 μm. The porosity of the composite layer lower region 63 was small at 4.9%. The thickness distribution coefficient of the composite layer 13 was 0.4. The C / Al ratio in the second region 62 was 0.20. In Comparative Example 1, the permeation flux was "unacceptable" and the separation performance was "excellent".

[0141] Comparing Examples 1 to 8 with Comparative Examples 1 to 5, in Comparative Examples 1, 4, and 5, the average thickness t ａｖｅ is larger than 1.0 μm, and in Comparative Examples 1 and 3 to 5, the standard deviation t ｓｄ In Comparative Example 2, the bond strength between the carbon film 12 and the support 11 is insufficient, causing the carbon film 12 to peel off. In contrast, in Examples 1 to 8, the average thickness t ａｖｅ is 0.01 μm or more and 1.0 μm or less, and the standard deviation t ｓｄ is 0.1 μm or more and 1.0 μm or less. As a result, in Examples 1 to 8, the permeation flux of the carbon membrane composite 1 can be increased. Specifically, the water permeation flux of the carbon membrane composite 1 can be increased to 0.4 kg / m ２ h or more. In addition, it is possible to realize a suitable separation performance of the carbon membrane composite 1. In Examples 1 to 8, the standard deviation t ｓｄ is 0.5 μm or more.

[0142] As described above, in Comparative Examples 1 and 3 to 5, the porosity of the composite layer lower region 63 is less than 5%. Furthermore, in Comparative Example 2, the bonding strength between the carbon membrane 12 and the support 11 is insufficient, causing the carbon membrane 12 to peel off. In contrast, in Examples 1 to 8, the porosity of the composite layer lower region 63 is 5% or more and 25% or less. This makes it possible to increase the permeation flux of the carbon membrane composite 1 in Examples 1 to 8. Specifically, the water permeation flux in the carbon membrane composite 1 is increased to 0.4 kg / m ２ h or more. In addition, the carbon membrane composite 1 can achieve a suitable separation performance.

[0143] Comparing Example 1 with Examples 2 to 4, the method of applying the non-solvent in step S12 was the immersion degassing method in Example 1, whereas a method other than the immersion degassing method (i.e., the immersion method or the flow-down method) was used in Examples 2 to 4. The permeation flux was "excellent" in Example 1, whereas it was "good" in Examples 2 to 4. From this, from the viewpoint of further increasing the permeation flux of the carbon membrane composite 1, the method of applying the non-solvent in step S12 is preferably the immersion degassing method.

[0144] Comparing Example 5 and Example 6, the thickness distribution coefficient of the composite layer 13 in Example 5 was 0.20 (i.e., 0.20 or more), whereas it was 0.18 (i.e., less than 0.20) in Example 6. The permeation flux was "excellent" in both Examples 5 and 6, but the water permeation flux in Example 5 was about 14% greater than the water permeation flux in Example 6. From this, from the viewpoint of further increasing the permeation flux of the carbon membrane composite 1, it is preferable that the thickness distribution coefficient of the composite layer 13 be 0.20 or more.

[0145] Comparing Example 5 and Example 7, the C / Al in the second region 62 is 0.09 (i.e., 0.05 or more) in Example 5, whereas it is 0.04 (i.e., less than 0.05) in Example 7. The separation performance was "excellent" in Example 5, whereas it was "good" in Example 7. From this, from the viewpoint of further improving the separation performance of the carbon membrane composite 1, it is preferable that the C / Al in the second region 62 is 0.05 or more.

[0146] Comparing Example 4 and Example 8, the C / Al in the second region 62 was 0.17 (i.e., 0.20 or less) in Example 4, whereas it was 0.25 (i.e., greater than 0.20) in Example 8. The separation performance was "excellent" in both Example 4 and Example 8, but the separation factor in Example 4 was about 12% greater than the separation factor in Example 8. From this, from the viewpoint of further improving the separation performance of the carbon membrane composite 1, it is preferable that the C / Al in the second region 62 be 0.20 or less.

[0147] Next, a second manufacturing method of the carbon membrane composite 1 will be described with reference to Fig. 12 and Fig. 13A to Fig. 13B. Fig. 12 is a diagram showing an example of a manufacturing flow of the carbon membrane composite 1 by the second manufacturing method. Fig. 13A to Fig. 13B are diagrams schematically showing a partial cross section of the carbon membrane composite 1 in the middle of manufacturing.

[0148] When the carbon membrane composite 1 is produced by the second production method, first, a support 11 is prepared (step S21) in the same manner as in step S11 (see FIG. 9 ) described above. Subsequently, a liquid in which the above-described precursor particles are dispersed in a dispersion medium (hereinafter also referred to as a "precursor dispersion") is applied to the support 11 from the side of the inner surface of the cell 111 (i.e., the inner surface 331 of the surface layer 33). The dispersion medium of the precursor dispersion flows out through the support 11, and the precursor particles remain on the inner surface 331 of the cell 111. As a result, a precursor particle layer 74, which is a layer of precursor particles stacked in the radial direction, is formed on the inner surface 331 of the surface layer 33, as shown in FIG. 13A (step S22).

[0149] The particle size of the precursor particles in the precursor dispersion is, for example, 100 nm to 500 nm. The dispersion medium of the precursor dispersion is a liquid in which the precursor particles do not dissolve, such as water. The precursor dispersion is a so-called suspension. Note that the particle size of the precursor particles in the precursor dispersion and the dispersion medium of the precursor dispersion are not limited to the above examples and may be changed in various ways.

[0150] The precursor dispersion liquid is applied in step S22 by, for example, a flow-down method, an immersion method, dip coating, spin coating, drip coating, spray coating, or filtration coating. In the present embodiment, the precursor dispersion liquid is applied by a flow-down method. Specifically, the support 11 is placed so that the longitudinal direction of the support 11 is approximately parallel to the direction of gravity, and the precursor dispersion liquid is poured into the upper opening of each cell 111, thereby applying the precursor dispersion liquid to the support 11.

[0151] When forming the precursor particle layer 74, if the thickness of the precursor particle layer 74 does not reach the desired thickness by applying the precursor dispersion only once, the precursor particle layer 74 may be formed by repeating the application of the precursor dispersion multiple times, as shown in FIG. 13A.

[0152] After step S22 is completed, the precursor particle layer 74 is melted to bond the particles together, thereby forming a precursor film 73, a thin film containing precursors, on the inner surface 331 of the surface layer 33 (i.e., the surface of the support 11), as shown in FIG. 13B. The precursor film 73 is a so-called polymer film. Furthermore, a portion of the molten precursor particle layer 74 penetrates from the inner surface 331 of the surface layer 33 into the interior of the surface layer 33 and spreads to the region near the inner surface 331 of the surface layer 33, forming a precursor layer 72 (step S23). The precursor layer 72 and the precursor film 73 are continuous. In FIG. 13B, the precursor layer 72 and the precursor film 73 are indicated by the same diagonal lines.

[0153] In step S23, for example, the precursor particle layer 74 is melted by heating. In this case, for example, the support 11 on which the precursor particle layer 74 has been formed is placed in an electric furnace, heated to 130°C to 150°C in an air atmosphere, and maintained at this temperature for 10 to 50 hours, thereby melting the precursor particle layer 74. Alternatively, in step S23, the precursor particle layer 74 may be melted by applying solvent vapor to the precursor particle layer 74. Examples of the solvent that can be used include organic solvents such as methanol, ethanol, acetone, tetrahydrofuran, NMP (N-methyl-2-pyrrolidone), and toluene. Note that the melting of the precursor particle layer 74 in step S23 may be performed by a combination of the above methods, or by a method other than those described above.

[0154] After step S23 is completed, the support 11 is heat-treated to harden the precursor film 73 on the surface of the support 11 and the precursor layer 72 inside the support 11. The heat treatment is performed, for example, by heating the support 11 to 90°C to 500°C using a dryer and holding the temperature for 0.5 hours to 60 hours.

[0155] When the curing of the precursor layer 72 and the precursor film 73 is completed, the support body 11 provided with the precursor layer 72 and the precursor film 73 is subjected to a carbonization treatment. As a result, a carbon film 12 (see FIG. 3 ) is formed from the precursor film 73 on the surface of the support body 11 (i.e., the inner surface 331 of the cell 111). Furthermore, a composite layer 13 (see FIG. 4 ) is formed from the precursor layer 72 in a region near the surface inside the support body 11. As a result, a carbon membrane composite 1 including the carbon film 12 and the composite layer 13 is obtained (step S24).

[0156] In the carbonization treatment in step S24, for example, the support 11 provided with the precursor layer 72 and the precursor film 73 is placed in an electric furnace and heated to 400°C to 1200°C. By setting the temperature during the carbonization treatment to 400°C or higher, the precursor film 73 is suitably carbonized, pores are formed, and the separation performance of the carbon film 12 is suitably exhibited. By setting the temperature during the carbonization treatment to 1200°C or lower, the carbon film 12 is prevented from becoming excessively dense, and the separation performance of the carbon film 12 is suitably exhibited. It is more preferable that the temperature during the carbonization treatment be 600°C to 900°C.

[0157] The carbonization treatment is preferably carried out in a non-oxidizing atmosphere. The non-oxidizing atmosphere means an atmosphere in which the precursor is not oxidized when carbonized within the temperature range. Specifically, the non-oxidizing atmosphere is an inert gas atmosphere such as nitrogen or argon, or a vacuum atmosphere.

[0158] In the second manufacturing method, the molten precursor particle layer 74 is allowed to soak into the support 11, thereby forming the precursor layer 72 near the surface of the support 11. Since the molten precursor particle layer 74 is less likely to soak into the support 11 than the precursor solution described above, in the second manufacturing method, the average thickness t ａｖｅ The thickness of the composite layer 13 can be easily reduced compared to the first manufacturing method using a precursor solution. ｓｄ The molten precursor particle layer 74 has a higher viscosity and is less likely to spread than the precursor solution described above, and therefore, in the second manufacturing method, the surface roughness of the carbon film 12 becomes larger than in the first manufacturing method using the precursor solution.

[0159] The surface roughness of the carbon film 12 is determined as follows. First, a polished cross section of a sample of the carbon membrane composite 1 is imaged using an SEM at a magnification of 5000 times or more. From the acquired image, a substantially rectangular region having a width of 2 μm along the surface of the support 11 (i.e., a substantially rectangular region having a width of 2 μm in the left-right direction in FIG. 14 ) is extracted as a cross-sectional image 91a illustrated in FIG. 14 . The cross-sectional image 91a includes at least a region near the surface 121 of the carbon film 12. In the example shown in FIG. 14 , the cross-sectional image 91a includes the carbon film 12, the boundary between the carbon film 12 and the surface layer 33 (i.e., the inner surface 331 of the surface layer 33), and a region near the inner surface 331 of the surface layer 33. The surface 121 of the carbon film 12 includes a plurality of convex portions 122 that protrude in a radial direction (i.e., the thickness direction of the carbon film 12) away from the surface layer 33, and a plurality of concave portions 123 that are recessed in the radial direction toward the surface layer 33.

[0160] Next, on the surface 121 of the carbon film 12, a straight line L3 tangent to one convex portion 122 that is radially farthest from the surface layer 33 and a straight line L4 tangent to one concave portion 123 that is closest to the surface layer 33 are drawn in the cross-sectional image 91a. The straight lines L3 and L4 are parallel to each other.

[0161] The line L3 is tangent to the surface 121 of the carbon film 12 only at the point of contact with the one protruding portion 122, and the other portions of the line L3 are spaced radially inward (i.e., upward in FIG. 14 ) from the carbon film 12. Alternatively, the line L3 may be tangent to not only the one protruding portion 122 but also another protruding portion 122. In this case, the other portions of the line L3 are spaced radially inward from the carbon film 12 except for the points of contact with the two protruding portions 122.

[0162] The straight line L4 is tangent to a space (hereinafter also referred to as the "intra-cell space 112") radially inward of the surface 121 of the carbon film 12 only at the point of contact with the one recess 123, and the other portions of the straight line L4 are spaced radially outward (i.e., downward in FIG. 14) from the intra-cell space 112. Alternatively, the straight line L4 may be tangent to not only the one recess 123 but also another recess 123. In this case, the other portions of the straight line L4 are spaced radially outward from the intra-cell space 112 except for the points of contact with these two recesses 123.

[0163] Next, the inclination of the parallel lines L3 and L4 is varied within a range that satisfies the above-described conditions for the lines L3 and L4, and the distance DL1 between the lines L3 and L4 in a direction perpendicular to the lines L3 and L4 is measured. The minimum value of the distance DL1 is then acquired as the provisional surface roughness in the cross-sectional image 91a. Note that when the distance DL1 is at its minimum, the line L3 is in contact with two convex portions 122 and / or the line L4 is in contact with two concave portions 123. The surface roughness of the carbon film 12 is calculated as the arithmetic mean of the provisional surface roughness in each of the 20 or more cross-sectional images 91a extracted from the above-described captured images.

[0164] Next, Examples 9 to 11 of the carbon membrane composite 1 produced by the second production method will be described with reference to Tables 4 to 6. Tables 5 and 6 also show the above-mentioned Comparative Examples 1, 2, and 4.

[0165]

[0166]

[0167]

[0168] The carbon membrane composites 1 of Examples 9 to 11 were produced by the production method shown in steps S21 to S24 described above. Examples 9 to 11 differed in the conditions for forming the precursor particle layer 74 in step S22 or for melting the precursor particle layer 74 in step S23. In Examples 9 to 11, the precursor layer 72 and precursor film 73 formed in step S23 were cured by heating the support 11 to 200°C in an air atmosphere using an electric furnace and maintaining the temperature for 5 hours. The carbonization treatment in step S24 was performed by heating the support 11 to 750°C in a vacuum atmosphere using a carbonization furnace and maintaining the temperature for 5 hours.

[0169] In Examples 9 to 11, the average pore size on the surface of the support 11, the thickness of the carbon membrane 12, the average thickness and standard deviation of the composite layer 13, the porosity of the composite layer lower region 63, and the thickness distribution coefficient of the composite layer 13 were obtained by the same methods as in Examples 1 to 8. The surface roughness of the carbon membrane 12 was obtained by the above-mentioned method (see FIG. 14 ). The permeation flux and separation performance in Table 6 were obtained and evaluated in the same way as in Table 3.

[0170] In Example 9, the average pore diameter on the surface of the support 11 was 0.07 μm. The dispersion medium of the precursor dispersion used in forming the precursor particle layer 74 in step S22 was water, and the particle size of the precursor particles was 200 nm. The number of times the precursor dispersion was applied to the support 11 in step S22 (i.e., the number of times the film was formed) was two. In step S23, the support 11 was placed in an electric furnace, heated to 140° C. in an air atmosphere, and maintained there for 50 hours, thereby melting the precursor particle layer 74 and forming the precursor layer 72 and the precursor film 73.

[0171] In Example 9, the thickness of the carbon film 12 was 0.62 μm, and the surface roughness of the carbon film 12 was 130 nm. ａｖｅ is 0.16 μm, and the standard deviation of the thickness of the composite layer 13 is t ｓｄwas 0.10 μm. The porosity of the composite layer lower region 63 was 11.8%. The thickness distribution coefficient of the composite layer 13 (i.e., (D50-D10) / (D90-D10)) was 0.35. The C / Al in the second region 62 was 0.08. In Example 9, the permeation flux was "excellent" and the separation performance was also "excellent".

[0172] Example 10 is similar to Example 9, except that the heating time of the precursor particle layer 74 in step S23 is 15 hours.

[0173] In Example 10, the thickness of the carbon film 12 was 0.62 μm, and the surface roughness of the carbon film 12 was 142 nm. ａｖｅ is 0.17 μm, and the standard deviation of the thickness of the composite layer 13 is t ｓｄ was 0.23 μm. The porosity of the composite layer lower region 63 was 9.9%. The thickness distribution coefficient of the composite layer 13 (i.e., (D50-D10) / (D90-D10)) was 0.39. The C / Al in the second region 62 was 0.08. In Example 10, the permeation flux was "excellent" and the separation performance was also "excellent".

[0174] Example 11 is the same as Example 10, except that the particle size of the precursor particles in the precursor dispersion in step S22 is 500 nm and the number of times the precursor dispersion is applied (i.e., the number of times the film is formed) is one.

[0175] In Example 11, the thickness of the carbon film 12 was 1.50 μm, and the surface roughness of the carbon film 12 was 330 nm. ａｖｅ is 0.30 μm, and the standard deviation of the thickness of the composite layer 13 is t ｓｄ The thickness distribution coefficient of the composite layer 13 (i.e., (D50-D10) / (D90-D10)) was 0.37. In Example 11, the permeation flux was "excellent" and the separation performance was "good."

[0176] Comparing Examples 9 to 11 with Comparative Examples 1, 3, and 4, in Comparative Examples 1 and 4, the average thickness t ａｖｅis larger than 1.0 μm, and in Comparative Examples 1, 3, and 4, the standard deviation t ｓｄ On the other hand, in Examples 9 to 11, the average thickness t ａｖｅ is 0.01 μm or more and 1.0 μm or less, and the standard deviation t ｓｄ is 0.1 μm or more and 1.0 μm or less. As a result, in Examples 9 to 11, the permeation flux of the carbon membrane composite 1 can be increased. Specifically, the water permeation flux of the carbon membrane composite 1 can be increased to 0.4 kg / m ２ h or more (more specifically, 1.2 kg / m ２ h or more). In addition, the carbon membrane composite 1 can achieve favorable separation performance.

[0177] As described above, in Comparative Examples 1, 3, and 4, the porosity of the composite layer lower region 63 is less than 5%. In contrast, in Examples 9 to 11, the porosity of the composite layer lower region 63 is 5% or more and 25% or less. This makes it possible to increase the permeation flux of the carbon membrane composite 1 in Examples 9 to 11. Specifically, the water permeation flux in the carbon membrane composite 1 is increased to 0.4 kg / m ２ h or more (more specifically, 1.2 kg / m ２ h or more). In addition, the carbon membrane composite 1 can achieve favorable separation performance.

[0178] As described above, in Comparative Examples 1, 3, and 4, the surface roughness of the carbon membrane 12 is less than 100 nm. In contrast, in Examples 9 to 11, the surface roughness of the carbon membrane 12 is 100 nm or more. This makes it possible to increase the permeation flux of the carbon membrane composite 1 in Examples 9 to 11. Specifically, the water permeation flux in the carbon membrane composite 1 is increased to 0.4 kg / m ２ h or more (more specifically, 1.2 kg / m ２ h or more).

[0179] Comparing Examples 9 and 10 with Example 11, the surface roughness of the carbon membrane 12 was 300 nm or less in Examples 9 and 10, whereas it was greater than 300 nm in Example 11. The separation performance was "excellent" in Examples 9 and 10, whereas it was "good" in Example 11. From this viewpoint, from the viewpoint of further improving the separation performance of the carbon membrane composite 1, it is preferable that the surface roughness of the carbon membrane 12 be 300 nm or less.

[0180] As explained above, the carbon membrane composite 1 includes a porous support 11 and a carbon membrane 12 provided on the surface of the support 11 (in the above example, the inner surface of the cell 111). Inside the support 11, a composite layer 13 is provided, which extends from the surface of the support 11 toward the inside of the support 11. In the composite layer 13, the voids 333 between the aggregate particles 332 of the support 11 are filled with carbon or carbide (i.e., at least one of carbon and carbide). In the thickness direction perpendicular to the surface of the support 11, the arithmetic mean of the thickness of the composite layer 13 (i.e., the average thickness t ａｖｅ ) is 0.01 μm or more and 1.0 μm or less, and the standard deviation t ｓｄ is 0.1 μm or more and 1.0 μm or less. In this way, in the carbon membrane composite 1, by reducing the thickness of the composite layer 13 and the variation in the thickness, it is possible to increase the permeation flux as described above. In addition, it is also possible to achieve favorable separation performance of the carbon membrane composite 1.

[0181] In the carbon membrane composite 1, in the thickness direction perpendicular to the surface of the support 11, a region obtained by excluding a region extending from the surface of the support 11 to a distance equal to the arithmetic mean of the thickness of the composite layer 13 (i.e., the first region 61) from a region extending from the surface of the support 11 to a distance four times the arithmetic mean of the thickness of the composite layer 13 (i.e., the second region 62) is defined as a composite layer lower region 63, and the porosity of the composite layer lower region 63 is 5% or more and 25% or less. This reduces the thickness and thickness variation of the composite layer 13 in the carbon membrane composite 1, in a manner similar to that described above. As a result, the permeation flux of the carbon membrane composite 1 can be increased, in a manner similar to that described above. Furthermore, favorable separation performance of the carbon membrane composite 1 can also be achieved.

[0182] As described above, the ratio of carbon to aluminum in the region extending from the surface of the support 11 to a depth four times the arithmetic mean of the thickness of the composite layer 13 in the thickness direction (i.e., C / Al in the second region 62) is preferably 0.05 or more and 0.2 or less. This makes it possible to preferably achieve both an increase in permeation flux and an improvement in separation performance of the carbon membrane composite 1, as described above.

[0183] As described above, it is preferable that (D50-D10) / (D90-D10) in the thickness distribution of the composite layer 13 (i.e., the thickness distribution coefficient of the composite layer 13) is 0.2 or more and 0.4 or less. This makes it possible to preferably achieve both an increase in permeation flux and an improvement in separation performance of the carbon membrane composite 1, as described above.

[0184] Preferably, the average pore diameter on the surface of the support 11 (in the above example, the inner surface 331 of the surface layer 33) is 0.01 μm or more. This makes it possible to easily manufacture the support 11 compared to when the average pore diameter is excessively small. As a result, it is also possible to easily manufacture the carbon membrane composite 1. Furthermore, it is also possible to suppress a decrease in the permeation performance of the support 11.

[0185] More preferably, the average pore size on the surface of the support 11 (in the above example, the inner surface 331 of the surface layer 33) is 0.07 μm or more. This further facilitates the production of the support 11 and the carbon membrane composite 1. It also makes it possible to further suppress the deterioration of the permeation performance of the support 11.

[0186] As described above, the support 11 preferably includes a first layer (i.e., the surface layer 33), a second layer (i.e., the intermediate layer 32), and a third layer (i.e., the substrate 31). The surface layer 33 extends from the surface of the support 11 (in the above example, the inner surface of the cell 111) toward the inside of the support 11. The intermediate layer 32 extends from the surface layer 33 toward the inside of the support 11. The intermediate layer 32 has a larger average pore diameter than the surface layer 33. The substrate 31 extends from the intermediate layer 32 toward the inside of the support 11. The substrate 31 has a larger average pore diameter than the intermediate layer 32. Furthermore, the average pore diameter of the substrate 31 is preferably 3 μm or more. This makes it possible to reduce the average pore diameter of the surface layer 33 of the support 11 (i.e., the average pore diameter on the surface of the support 11) while suppressing a decrease in the permeability performance of the support 11.

[0187] As described above, the thickness of the carbon membrane 12 in the thickness direction is preferably 0.1 μm or more and 5 μm or less, which makes it possible to preferably achieve both an increase in permeation flux and an improvement in separation performance of the carbon membrane composite 1.

[0188] As described above, the surface roughness of the carbon membrane 12 is preferably 100 nm or more and 300 nm or less, which can further increase the permeation flux of the carbon membrane composite 1 and further improve the separation performance of the carbon membrane composite 1.

[0189] The first method for producing the carbon membrane composite 1 described above includes the steps of: applying a non-solvent to a porous support 11 to allow the non-solvent to penetrate the support 11 and form a non-solvent layer 71 inside the support 11 (step S12); removing a portion of the non-solvent layer 71 from the surface side of the support 11 (in the above example, the inner surface of the cell 111) (step S13); applying a precursor solution containing a precursor from the surface side of the support 11 while applying pressure from the opposite side of the support 11 to form a precursor membrane 73, which is a membrane containing the precursor, on the surface of the support 11 (step S14); drying the precursor membrane 73 and the non-solvent layer 71 while applying pressure from the opposite side of the support 11 to the surface (step S15); and carbonizing the precursor membrane 73 to form a carbon membrane 12 on the surface of the support 11 (step S16). This allows for the production of a carbon membrane composite 1 with reduced thickness and variation in the composite layer 13. The carbon membrane composite 1 can increase the permeation flux as described above. Furthermore, it can also achieve favorable separation performance.

[0190] The second method for producing the carbon membrane composite 1 described above includes the steps of: applying a liquid (i.e., precursor dispersion liquid) in which precursor particles are dispersed to the surface side of the support 11 to form a layer of precursor particles (i.e., precursor particle layer 74) on the surface of the support 11 (step S22); melting the precursor particle layer 74 to form a precursor membrane 73, which is a membrane containing precursors, on the surface of the support 11 (step S23); and carbonizing the precursor membrane 73 to form a carbon membrane 12 on the surface of the support 11 (step S24). This makes it possible to suitably produce a carbon membrane composite 1 in which the thickness of the composite layer 13 and the variation in thickness are reduced. As described above, the carbon membrane composite 1 can increase the permeation flux. Furthermore, favorable separation performance can also be achieved.

[0191] The above-described carbon membrane composite 1 and the method for manufacturing the carbon membrane composite 1 can be modified in various ways.

[0192] For example, the thickness of the carbon film 12 may be less than 0.1 μm or more than 5 μm, and the surface roughness of the carbon film 12 may be less than 100 nm or more than 300 nm.

[0193] The substrate 31 may have an average pore size of less than 3 μm.

[0194] The average pore size on the surface of the support 11 may be less than 0.07 μm, or may be less than 0.01 μm.

[0195] The thickness distribution coefficient of the composite layer 13 (ie, (D50-D10) / (D90-D10)) may be less than 0.2 and may be greater than 0.4.

[0196] The C / Al ratio in the second region 62 may be less than 0.05 or greater than 0.2.

[0197] In the carbon membrane composite 1, when the porosity of the composite layer lower region 63 is 5% or more and 25% or less, the average thickness t ａｖｅ For example, the average thickness t of the composite layer 13 may be ａｖｅ In addition, if the porosity of the composite layer lower region 63 is 5% or more and 25% or less, the standard deviation t ｓｄ may be less than 0.1 μm and may be greater than 1.0 μm.

[0198] In the carbon membrane composite 1, the average thickness t ａｖｅ is 0.01 μm or more and 1.0 μm or less, and the standard deviation t ｓｄ However, if the porosity is greater than or equal to 0.1 μm and less than or equal to 1.0 μm, the porosity of the composite layer lower region 63 may be less than 5% and greater than 25%.

[0199] The support 11 (i.e., the base material 31, the intermediate layer 32, and the surface layer 33) does not necessarily have to be made of ceramics, but may be made of other materials such as metal.

[0200] In the support 11 of the carbon membrane composite 1, either the surface layer 33 or the intermediate layer 32 may be omitted, or both the surface layer 33 and the intermediate layer 32 may be omitted. When both the surface layer 33 and the intermediate layer 32 are omitted, the substrate 31 functions alone as the support 11, and the carbon membrane 12 is provided directly on the surface of the substrate 31. Alternatively, in the carbon membrane composite 1, in addition to the surface layer 33 and the intermediate layer 32, another layer having an average pore diameter smaller than that of the substrate 31 may be provided directly or indirectly on the substrate 31. The other layer is laminated on the substrate 31 together with the surface layer 33 and the intermediate layer 32.

[0201] The partial removal of the non-solvent layer 71 in the above-mentioned step S13 does not necessarily have to be performed by a drying process using air blowing, but may be performed by a drying process other than air blowing, or may be performed by a method other than a drying process.

[0202] The curing of the precursor film 73 and the precursor layer 72 between the above-described steps S15 and S16 does not necessarily have to be performed and may be omitted.

[0203] The method for producing the carbon membrane composite 1 is not limited to the above, and the carbon membrane composite 1 may be produced by other production methods.

[0204] The configurations in the above-described embodiment and each modification may be combined as appropriate as long as they are not mutually contradictory.

[0205] While the invention has been particularly illustrated and described, it should be understood that the foregoing description is illustrative and not restrictive, and that numerous modifications and variations are possible without departing from the scope of the invention.

[0206] The present invention can be used for separating various substances using a carbon membrane.

[0207] REFERENCE SIGNS LIST 1 carbon membrane composite 11 support 12 carbon membrane 13 composite layer 31 substrate 32 intermediate layer 33 surface layer 61 first region 62 second region 63 composite layer lower region 71 non-solvent layer 72 precursor layer 73 precursor membrane 74 precursor particle layer 331 inner surface 333 voids S11 to S16, S21 to S24 steps

Claims

1. A carbon membrane composite comprising: a porous support; and a carbon membrane provided on the surface of the support; wherein a composite layer is provided inside the support, the composite layer extending from the surface of the support toward the inside of the support and having voids between aggregate particles of the support filled with carbon or carbide; and wherein the arithmetic mean thickness of the composite layer in a thickness direction perpendicular to the surface of the support is 0.01 μm or more and 1.0 μm or less, and the standard deviation of the thickness of the composite layer is 0.1 μm or more and 1.0 μm or less.

2. A carbon membrane composite comprising: a porous support; and a carbon membrane provided on the surface of the support; wherein a composite layer is provided inside the support, the composite layer extending from the surface of the support toward the inside of the support and having voids between aggregate particles of the support filled with carbon or carbide; and wherein a lower region of the composite layer, in a thickness direction perpendicular to the surface of the support, is defined as a region extending from the surface of the support to a distance four times the arithmetic mean of the thickness of the composite layer, excluding a region extending from the surface of the support to a distance equal to the arithmetic mean of the thickness of the composite layer, and the porosity of the lower region of the composite layer is 5% or more and 25% or less.

3. A carbon membrane composite according to claim 2, wherein the ratio of carbon to aluminum in a region extending from the surface of the support to a depth four times the arithmetic mean of the thickness of the composite layer in the thickness direction is 0.05 or more and 0.2 or less.

4. A carbon membrane composite according to any one of claims 1 to 3, wherein (D50-D10) / (D90-D10) in the thickness distribution of the composite layer is 0.2 or more and 0.4 or less.

5. A carbon membrane composite according to any one of claims 1 to 3, wherein the average pore size on the surface of the support is 0.01 µm or more.

6. A carbon membrane composite according to claim 5, wherein the average pore size on the surface of the support is 0.07 μm or more.

7. A carbon membrane composite according to any one of claims 1 to 3, wherein the support comprises: a first layer extending from the surface of the support toward the inside of the support; a second layer extending from the first layer toward the inside of the support and having a larger average pore diameter than the first layer; and a third layer extending from the second layer toward the inside of the support and having a larger average pore diameter than the second layer, wherein the average pore diameter of the third layer is 3 μm or more.

8. A carbon membrane composite according to any one of claims 1 to 3, wherein the thickness of the carbon membrane in the thickness direction is 0.1 µm or more and 5 µm or less.

9. A carbon membrane composite according to any one of claims 1 to 3, wherein the surface roughness of the carbon membrane is 100 nm or more and 300 nm or less.

10. A method for producing a carbon membrane composite, comprising: a) applying a non-solvent to a porous support to allow the non-solvent to penetrate the support, thereby forming a non-solvent layer inside the support; b) removing a portion of the non-solvent layer from the surface side of the support; c) applying a precursor solution containing a precursor from the surface side of the support while applying pressure from the side opposite to the surface of the support, thereby forming a precursor film that is a film containing the precursor on the surface of the support; d) drying the precursor film and the non-solvent layer while applying pressure from the side opposite to the surface of the support; and e) carbonizing the precursor film to form a carbon film on the surface of the support.

11. A method for producing a carbon membrane composite, comprising: a) applying a liquid having precursor particles dispersed therein from the surface side of the support to form a layer of precursor particles on the surface of the support; b) melting the layer of precursor particles to form a precursor film, which is a film containing the precursor, on the surface of the support; and c) carbonizing the precursor film to form a carbon film on the surface of the support.

Images