CARBON FIBER BUNDLE AND METHOD FOR MANUFACTURING SAME

Abstract

 The carbon fiber bundle of the present invention is a carbon fiber bundle that satisfies a predetermined elastic modulus of strands, a predetermined tensile strength of strands, a predetermined knot strength, and a predetermined average single fiber diameter, in which the probability that a flaw with a size of 50 nm or more exists on the collected fracture surface when performing a single fiber tensile test with a gauge length of 10 mm is a predetermined ratio. Such a carbon fiber bundle is suitably obtained by filtering a spinning dope solution in which a polyacrylonitrile copolymer is dissolved in a solvent, at a predetermined filtration speed, using a filter medium having a predetermined particle retention and a filter basis weight, then spinning the filtered spinning dope solution to obtain a precursor fiber bundle for carbon fiber, and heat-treating the obtained precursor fiber bundle for carbon fiber at an appropriate temperature profile in an oxidizing atmosphere until reaching a predetermined density to obtain an oxidized fiber bundle, and then heat-treating the oxidized fiber bundle at a predetermined temperature in an inert atmosphere. An object of the present invention is to provide a carbon fiber bundle that exhibits excellent tensile strength of strands and elastic modulus of strands in a well-balanced manner and also has excellent knot strength, and a method for manufacturing the same.

Description

TECHNICAL FIELD

[0001] The present invention relates to a carbon fiber bundle and a method for manufacturing the same.

BACKGROUND ART

[0002] Carbon fiber bundles have been widely used as reinforcing fibers for composite materials, and there is a strong demand for more performance. In particular, in order to reduce the weight of members such as pressure vessels, it is required to improve mechanical properties such as tensile strength of resin-impregnated strands and elastic modulus of resin-impregnated strands of the carbon fiber bundle (hereinafter simply referred to as tensile strength of strands and elastic modulus of strands) in a well-balanced manner. At the same time, it is necessary to reduce environmental load in the manufacture of carbon fiber bundles. Generally, a polyacrylonitrile-based carbon fiber bundle is obtained through a process in which a precursor fiber bundle for carbon fiber is heat-treated in an oxidizing atmosphere of 200 to 300°C (oxidation process) and then heat-treated in an inert atmosphere of 1000°C or more (carbonization process). At that time, since carbon, nitrogen and hydrogen atoms contained in polyacrylonitrile are desorbed by thermal degradation, the yield of the carbon fiber bundles (hereinafter also referred to as carbonization yield) is about half. It is necessary to increase the yield of carbon fiber bundles with the same manufacturing energy from the viewpoint of reducing the manufacturing energy per production amount, that is, environmental load.

[0003] For this reason, so far, many techniques have been proposed for the purpose of improving the tensile strength of strands or carbonization yield of carbon fiber bundles by optimizing the oxidizing conditions (Patent Documents 1 to 5).

[0004] In Patent Document 1, studies have been made to improve the tensile strength of strands of the carbon fiber bundle by minimizing the amount of heat (J · h/g) given by high-temperature treatment in the oxidation process. In Patent Document 2, it was proposed to set the oxidation temperature to a high temperature according to the amount of oxygen added in the middle of the oxidation process, and in Patent Document 3, it was proposed to oxidize at a temperature as high as possible by repeating heating and cooling so that the precursor fiber bundle for carbon fiber does not thermally run away, in order to shorten the oxidation process. Moreover, Patent Documents 4 and 5 proposed attempts to increase the carbonization yield by increasing the density of oxidized fiber bundle in a short time by heating the precursor fiber bundle for carbon fiber in an oxidizing atmosphere in an initial stage of oxidation, and then bringing it into contact with a high-temperature heating roller at 250 to 300°C.

[0005] Patent Documents 6 and 7 have proposed carbon fiber bundles with high knot strength that reflect mechanical properties in a direction other than a fiber axis direction and exhibit sufficient mechanical properties in pseudo-isotropic materials.

[0006] Patent Document 8 has proposed a carbon fiber bundle that exhibits a high carbonization yield, excellent tensile strength of strands and elastic modulus of strands in a well-balanced manner, and further satisfies excellent knot strength at the same time since an oxidized fiber bundle having a specific density can be obtained by performing high-temperature heat treatment in the latter half at an appropriate temperature profile in the oxidation process when obtaining an oxidized fiber bundle having a specific density in order to satisfy a high carbonization yield.

[0007] On the other hand, the carbon fiber is a brittle material, and since a slight surface flaw and an internal flaw cause a decrease in tensile strength of strands, delicate attention has been paid to the generation of flaw. For example, Patent Document 9 has proposed to reduce flaws on the surface of the carbon fiber to obtain a carbon fiber bundle having a high tensile strength of strands, by densification of the precursor fiber bundle for carbon fiber, reduction of dust during the manufacturing process, and removal of flaw by electrochemical treatment.

PRIOR ART DOCUMENTS

PATENT DOCUMENTS

[0008] 

Patent Document 1: Japanese Patent Laid-open Publication No. 2012-82541

Patent Document 2: Japanese Patent Laid-open Publication No. S58-163729

Patent Document 3: Japanese Patent Laid-open Publication No. H6-294020

Patent Document 4: Japanese Patent Laid-open Publication No. 2013-23778

Patent Document 5: Japanese Patent Laid-open Publication No. 2014-74242

Patent Document 6: WO 2013/157613 A

Patent Document 7: Japanese Patent Laid-open Publication No. 2015-096664

Patent Document 8: Japanese Patent Laid-open Publication No. 2017-66580

Patent Document 9: Japanese Examined Patent Publication No. H8-6210

SUMMARY OF THE INVENTION

PROBLEMS TO BE SOLVED BY THE INVENTION

[0009] However, in the proposal of Patent Document 1, an attempt has been made to reduce the integrated value of the amount of heat given in the oxidation process, which is not sufficient for achieving both the tensile strength of strands and the carbonization yield. In addition, in the proposals of Patent Documents 2 and 3, since the oxidation temperature has been increased and the oxidation time has been shortened, the oxidation temperature control that can satisfy the required tensile strength of strands is not performed, and suppression of stress concentration on a difference between skin-core structure has been a problem. Moreover, in the proposals of Patent Documents 4 and 5, while heat treatment has been performed at a high temperature using a heating roller with high heat transfer efficiency in order to perform heat treatment at a high temperature in a short time in the latter half of the oxidation process, sufficient tensile strength of strands has not been obtained due to too short heat treatment time at a high temperature, and generation of flaw due to adhesion between single fibers when passing through the roller. Although the proposal of Patent Document 6 states that the knot strength is increased by adjusting the oxidation process even when the single fiber diameter is large, the effect is limited due to the structure distribution in the single fiber at the time of oxidation, and the level of knot strength has been insufficient. Although the proposal of Patent Document 7 states that the knot strength is increased by mainly adjusting the surface treatment of the carbon fiber bundle and the sizing agent, it is limited to those having a low single fiber diameter, and in the case of having a low single fiber diameter, the fracture tension of the single fiber is lowered during the manufacturing process, so that there is a problem that the quality of the manufacturing process is lowered due to fiber fracture. In the proposal of Patent Document 8, the tensile strength of strands and knot strength have been increased by high-temperature heat treatment in the latter half at an appropriate temperature profile in the oxidation process, but the control of flaws affecting these characteristics has been not sufficient, and there has been room for improvement. In Patent Document 9, although the flaws on the carbon fiber surface can be effectively removed by electrochemical treatment, strong electrochemical treatment is required to remove the flaws, and a long electrochemical treatment bath is required, so that there has been a problem that it is difficult to implement industrially. In addition, there has been also a problem that a brittle layer that may lead to deterioration of mechanical properties of a composite is formed on the surface of the carbon fiber by strong electrochemical treatment. Furthermore, as a flaw, characteristics of flaws in the fracture surface collected when the single fiber tensile test was performed with a gauge length of 50 mm are defined. However, the gauge length that affects the tensile strength of strands and the tensile strength of the composite material is shorter than 10 mm, so that there has been also an essential problem that the carbon fiber bundles that increase the tensile strength of the composite material are not necessarily obtained simply by defining the characteristics of flaws observed at a gauge length of 50 mm.

[0010] An object of the present invention is to provide a method for manufacturing a carbon fiber bundle that exhibits the tensile strength of strands and the elastic modulus of strands in a well-balanced manner and has excellent knot strength without impairing productivity in order to solve the above-described problems in the prior art.

SOLUTIONS TO THE PROBLEMS

[0011] In order to achieve the above object, the method for manufacturing a carbon fiber bundle of the present invention is a method including filtering a spinning dope solution in which a polyacrylonitrile copolymer is dissolved in a solvent, using a filter medium having a particle retention B (µm) and a filter basis weight D (g/m²), under conditions where a filtration speed A (cm/hour) satisfies following equations (1) to (3), then spinning the filtered spinning dope solution to obtain a precursor fiber bundle for carbon fiber,





heat-treating the obtained precursor fiber bundle for carbon fiber in an oxidizing atmosphere until a density reaches 1.32 to 1.35 g/cm³, then heat-treating at 275°C or more and 295°C or less in an oxidizing atmosphere until a density reaches 1.46 to 1.50 g/cm³ to obtain an oxidized fiber bundle, and then heat-treating the oxidized fiber bundle at 1200 to 1800°C in an inert atmosphere.

[0012] Also, the carbon fiber bundle of the present invention is a carbon fiber bundle having an elastic modulus of strands of 240 to 280 GPa, a tensile strength of strands of 5.8 GPa or more, a knot strength K [MPa] of -88d + 1390 ≤ K (d: average single fiber diameter [µm]), and an average single fiber diameter in the range of 6.5 to 8.0 µm, wherein a probability that a flaw with a size of 50 nm or more exists on a fracture surface, which is collected when a single fiber tensile test is performed with a gauge length of 10 mm, is 35% or less.

EFFECTS OF THE INVENTION

[0013] According to the method of the present invention, when obtaining an oxidized fiber bundle, an oxidized fiber bundle having a specific density can be obtained by heat-treating at an appropriate temperature profile in the oxidation process, whereby flaws governing the tensile strength of strands and the knot strength are controlled to be very small, and thus a carbon fiber bundle that exhibits the tensile strength of strands and the elastic modulus of strands in a well-balanced manner and has excellent knot strength can be manufactured without impairing productivity. Moreover, according to the carbon fiber bundle of the present invention, it becomes a carbon fiber bundle which satisfies the productivity at the time of manufacturing a composite material.

BRIEF DESCRIPTION OF THE DRAWINGS

[0014] 

Fig. 1 is a scanning electron microscope (SEM) image of a fracture surface of a carbon fiber. Radial streaks converging to one point are confirmed.

Fig. 2 is an enlarged image of the vicinity of a fracture origin in Fig. 1. Flaws like attached substances are confirmed.

Fig. 3 is an enlarged image of the vicinity of a fracture origin of another fracture surface. Flaws like dent are confirmed.

Fig. 4 is an enlarged image of the vicinity of a fracture origin of another fracture surface. No noticeable morphological features of 50 nm or more are confirmed.

EMBODIMENTS OF THE INVENTION

[0015] The carbon fiber bundle of the present invention has a tensile strength of strands of 5.8 GPa or more, and preferably 6.0 GPa or more. In a tensile strength of strands of 5.8 GPa or more, when a composite material is manufactured using the carbon fiber bundle, the composite material exhibits good tensile strength. The higher tensile strength of strands of the carbon fiber bundle is better. However, even when the tensile strength of strands is 7.0 GPa or less, a sufficient tensile strength of the composite material can be obtained. The tensile strength of strands can be determined by a method described in the strand tensile test of the carbon fiber bundle described later. In addition, this tensile strength of strands can be controlled by using the method for manufacturing a carbon fiber bundle of the present invention described later.

[0016] The carbon fiber bundle of the present invention has an elastic modulus of strands of 240 to 280 GPa, preferably 245 to 275 GPa, and more preferably 250 to 270 GPa. An elastic modulus of strands of 240 to 280 GPa is preferable because of excellent balance between the elastic modulus of strands and the tensile strength of strands. In particular, by controlling the elastic modulus of strands to 250 to 270 GPa, a carbon fiber bundle having excellent tensile strength of strands is easily obtained. The elastic modulus of strands can be determined by a method described in the strand tensile test of the carbon fiber bundle described later. At this time, the strain range is set to 0.1 to 0.6%. The elastic modulus of strands of the carbon fiber bundle can be controlled mainly by applying tension to the fiber bundle in any of the heat treatment processes in the manufacturing process of the carbon fiber bundle, improving the difference between skin-core structure that is the structure distribution within the single fiber, or changing the carbonization temperature.

[0017] Also, in the carbon fiber bundle of the present invention, a knot strength K obtained by forming a knot part at the midpoint portion of the carbon fiber bundle and performing a fiber bundle tensile test is preferably 700 MPa or more, more preferably 740 MPa or more, and further preferably 770 MPa or more. The knot strength can be obtained by a method described in the knot strength of the carbon fiber bundle described later. The knot strength is an index that reflects mechanical properties of the fiber bundle in a direction other than a fiber axis direction. When manufacturing a composite material, a force in a bending direction is loaded on the carbon fiber bundle. When the number of filaments is increased in order to manufacture the composite material efficiently, it is difficult to increase the running speed of the fiber bundle during manufacture of the composite material due to generation of fuzz. However, when having a knot strength of 700 MPa or more, even in conditions where the running speed of the fiber bundle is high, a composite material with high quality can be obtained. In order to increase the knot strength of the carbon fiber bundle, in the method for manufacturing a carbon fiber bundle of the present invention described later, it is particularly preferable to control so that structural parameters in an oxidation process and a pre-carbonization process are within preferable ranges. Further, the knot strength can be also increased by reducing flaws on the surface of the carbon fiber.

[0018] The carbon fiber bundle preferably has the number of filaments of 10,000 to 60,000. When the number of filaments is 10,000 or more, a composite material can be manufactured with high productivity. When the number of filaments is 60,000 or less, the generation of fuzz at the time of manufacturing a composite material can be suppressed, and the running speed of the fiber bundle is increased, so that the productivity is easily increased.

[0019] Moreover, the carbon fiber bundle has a knot strength K [MPa] (= N/mm²) of -88d + 1390 ≤ K (where d is an average single fiber diameter [µm]). It is preferable that the carbon fiber bundle satisfies a relational expression -88d + 1410 ≤ K. Such a relational expression indicates that the knot strength is high for the average single fiber diameter. When the knot strength K satisfies -88d + 1390 ≤ K, during a filament winding molding process, even in a carbon fiber bundle with a large average single fiber diameter which is prone to fuzz due to abrasion with guide parts or rollers, it is possible to suppress the generation of fuzz and mold by increasing the running speed of the fiber bundle. In order to satisfy this relational expression, it is preferable to appropriately set the oxidizing conditions according to the average single fiber diameter by the manufacturing method of the present invention described later.

[0020] In the carbon fiber bundle, the probability that a flaw with a size of 50 nm or more exists on the fracture surface, which is collected when a single fiber tensile test is performed with a gauge length of 10 mm, is preferably 35% or less, more preferably 30% or less, and further preferably 25% or less. It is known that the tensile fracture of carbon fiber starts from a flaw. It is known that there are various types of flaws to be fracture origins of carbon fibers, such as voids, damages on the fiber surface, dents, attached substances, or adhesion marks that remain after single fibers adhere to each other by the heat of heat treatment. However, in the present invention, morphological features that can be observed by scanning electron microscope (SEM) observation are collectively referred to as "flaws" without particularly distinguishing all of them. As a result of studies by the present inventors, it has been found that the tensile strength of strands of the carbon fiber bundle is greatly increased when the probability that a flaw with a size of 50 nm or more exists on the fracture surface, which is collected when a single fiber tensile test is performed with a gauge length of 10 mm, is set to 35% or less. What is important here is that the gauge length is set to 10 mm. When a single fiber tensile test was performed with a longer gauge length, for example, a gauge length of 50 mm, even if the probability that a flaw of a certain size or larger exists was examined as described above, it has been found that the probability is not necessarily correlated with the tensile strength of strands and the tensile strength of the composite material as a result of the studies by the present inventors. The reason why it is effective to set the gauge length to 10 mm is considered that the gauge length that governs the tensile strength of strands and the tensile strength of the composite material (generally referred to as effective gauge length) is shorter than 10 mm. The probability that a flaw with a size of 50 nm or more exists on the fracture surface, which is collected when a single fiber tensile test is performed with a gauge length of 10 mm, is set to 35% or less, whereby flaws affecting the tensile strength of strands of the carbon fiber bundle and the tensile strength of the composite material are effectively reduced, as a result, the tensile strength of strands and the tensile strength of the composite material reach high levels. "The probability that a flaw with a size of 50 nm or more exists on the fracture surface which is collected when a single fiber tensile test is performed with a gauge length of 10 mm" is reduced by controlling filtration conditions of a spinning dope solution, that are filtration speed, particle retention, and filter basis weight, according to the methods described later, and effectively removing foreign substances in the spinning dope solution.

[0021] In the carbon fiber bundle of the present invention, the average single fiber diameter is 6.5 to 8.0 µm, preferably 6.7 to 8.0 µm, more preferably 7.0 to 8.0 µm, further preferably 7.3 to 8.0 µm, and most preferably 7.5 to 8.0 µm. As the average single fiber diameter is smaller, the difference between skin-core structure tends to decrease. However, when a composite material is prepared, it may cause insufficient impregnation due to a high matrix resin viscosity, which may lower the tensile strength of the composite material. An average single fiber diameter of 6.5 to 8.0 µm is preferred because insufficient impregnation of a matrix resin is unlikely to occur and a high carbonization yield and tensile strength of strands are stably exhibited. The average single fiber diameter can be calculated from the mass and density per unit length of the carbon fiber bundle and the number of filaments. The average single fiber diameter of the carbon fiber bundle is increased by increasing the average single fiber diameter of the precursor fiber bundle for carbon fiber, increasing the carbonization yield in the carbonization process by controlling the oxidizing conditions, and lowering the pre-carbonization stretch ratio.

[0022] The carbon fiber bundle preferably has a mean surface roughness Ra of a single fiber surface measured by an atomic force microscope (AFM) of 1.8 nm or less. Details of the measurement method will be described later. The mean surface roughness of the precursor fiber bundle for carbon fiber is substantially maintained even in the carbon fiber bundle. The mean surface roughness is preferably 1.0 to 1.8 nm, and further preferably 1.6 nm or less. When the mean surface roughness exceeds 1.8 nm, it tends to be a stress concentration point during tension, and the tensile strength of strands may decrease. The lower mean surface roughness is better. However, when the mean surface roughness is less than 1.0 nm, the effect is often almost saturated. The mean surface roughness of the carbon fiber bundle can be controlled by appropriately controlling spinning conditions of the precursor fiber bundle for carbon fiber (spinning method and coagulation bath conditions) and reducing the surface flaws of the carbon fiber bundle.

[0023] The carbon fiber bundle has an area ratio (hereinafter referred to as the skin layer ratio) in a cross section of a blackened thickness of an outer peripheral portion of the cross section perpendicular to the fiber axis direction of the single fiber of the carbon fiber of preferably 90% by area or more, more preferably 90 to 95% by area, and further preferably 90 to 93% by area. Here, the skin layer ratio is an area ratio (%) obtained by dividing an area occupied by the blackened thickness seen in the outer peripheral portion when observing a cross section perpendicular to the fiber axis direction of the single fiber of the carbon fiber with an optical microscope, by an entire cross-sectional area. Since the degree of orientation of the crystal part is low and the elastic modulus of strands is low in the inside of the blackened thickness of the single fiber of the carbon fiber, the higher the skin layer ratio, the more the surface layer stress concentration can be suppressed, so that high tensile strength of strands can be exhibited. When the skin layer ratio is low, a high carbonization yield and a high tensile strength of strands are hardly exhibited. When the skin layer ratio is 90% by area or more, the ratio of stress-bearing part on the outer peripheral portion is sufficiently large, so that stress concentration on the surface layer is suppressed. When the skin layer ratio exceeds 95% by area, the effect of suppressing stress concentration on the surface layer is saturated, on the other hand, the tensile strength of strands may decrease due to deviation of the oxidization temperature from optimum temperature. The blackened thickness can be measured by embedding a carbon fiber bundle in a resin, polishing a cross section perpendicular to the fiber axis direction, and observing the cross section with an optical microscope. Details will be described later.

[0024] In order to solve the problems of the present invention, the method for manufacturing a carbon fiber bundle of the present invention is based on the fact that it has been found that a carbon fiber bundle in which the number of flaws governing the tensile strength of strands and the knot strength is controlled to be extremely small, and a high carbonization yield and excellent tensile strength of strands and knot strength are exhibited is obtained by performing high-temperature heat treatment in the latter half at an appropriate temperature profile in the oxidation process so that the oxidized fiber bundle has a specific density. A preferred embodiment for carrying out the present invention will be described in detail below.

[0025] The precursor fiber bundle for carbon fiber can be obtained by spinning a spinning dope solution in which a polyacrylonitrile copolymer is dissolved in a solvent. At this time, the spinning dope solution is filtered under specific conditions to effectively remove foreign substances in the spinning dope solution, and then the filtered spinning dope solution is spun to obtain a precursor fiber bundle for carbon fiber. The obtained precursor fiber bundle for carbon fiber is subjected to at least an oxidation process, a pre-carbonization process and a carbonization process, so that a carbon fiber bundle having a high tensile strength of strands with few flaws can be obtained. As the polyacrylonitrile copolymer, it is preferable to use a polyacrylonitrile copolymer containing other monomers in addition to acrylonitrile as a main component. Specifically, the polyacrylonitrile copolymer preferably contains 90 to 100% by mass of acrylonitrile and less than 10% by mass of a copolymerizable monomer.

[0026] The polyacrylonitrile copolymer preferably contains a copolymer component such as itaconic acid, acrylamide and methacrylic acid, from the viewpoint of improving the stability of the spinning process, the viewpoint of efficiently performing the oxidation treatment, and the like.

[0027] The method for manufacturing the polyacrylonitrile copolymer can be selected from known polymerization methods. In the manufacture of a precursor fiber bundle for carbon fiber, the spinning dope solution is a solution prepared by dissolving the polyacrylonitrile copolymer in a solvent in which polyacrylonitrile such as dimethyl sulfoxide, dimethylformamide, dimethylacetamide or nitric acid/zinc chloride/aqueous sodium rhodanide solution is soluble.

[0028] Prior to spinning the spinning dope solution as described above, it is preferable to pass the spinning dope solution through a filter device to remove impurities mixed in the polymer raw material and each process. Here, the filter device means a facility for filtering and removing foreign substances present in the spinning dope solution, composed of an inflow path for introducing the spinning dope solution to be subjected to filtration into the filter device, a filter medium for filtering the spinning dope solution, an outflow path for guiding the filtered spinning dope solution to outside the filter device, and a container for storing these. Here, the filter medium is a means for filtering a spinning dope solution stored in the filter device.

[0029] As the form of the filter medium, a leaf disc type filter, a candle type filter, a pleated candle type filter or the like is used. The filter medium of the candle type filter or pleated candle type filter has constant curvature, whereas the leaf disc type filter can use the filter medium in a substantially planar form, so that this is preferable because it has an advantage that the pore diameter distribution hardly spreads and the cleaning property is easily maintained.

[0030] The filter medium is a member that plays a direct role for removing foreign substances present in the spinning dope solution. The filter medium is required to hold the determined opening diameters with narrow variations, and additionally, chemical stability to a substance to be treated, heat resistance and pressure resistance are required. As such a filter medium, a wire gauze prepared by weaving metal fibers, glass nonwoven fabric, filter medium made of a sintered metal fiber tissue and the like are preferably used. The material of the filter medium is not particularly limited as long as it is inert to the spinning dope solution and contains no elutable component into the solvent, but metals are more preferable from the viewpoint of durability and cost. As the specific metals, in addition to stainless steel (SUS304, SUS304L, SUS316, SUS316L, etc.), "INCONEL" (registered trademark) and "HASTELLOY" (registered trademark), various alloys based on nickel, titanium and cobalt are selected. Methods for manufacturing metal fibers include so-called bundle drawing, in which a large number of wires are collected as a bundle and the diameter is reduced by drawing, then individual wires are separated to reduce the diameter, coiled sheet shaving, chatter vibration shaving, and the like. In a case where the filter medium is wire gauze, because the metal fibers are necessary to be of not fiber bundles but single fibers, it is manufactured by a method including repeating wire drawing and heat treatment, or the like.

[0031] In filtration of the spinning dope solution, as the opening of the filter medium is smaller, foreign substances in the spinning dope solution are easily removed, but clogging of the filter medium more frequently occurs. In the present invention, as the removal performance for foreign substances, a "particle retention" is used. Here, the particle retention (µm) is a particle diameter (diameter) of spherical particles, 95% or more of which are collected when the particles pass through the filter medium. The particle retention can be measured according to a method of JIS standard (JIS-B8356-8: 2002). The fact that the particle retention is low and the fact that the particle retention is excellent are synonymous. In addition, as the filter thickness becomes thicker, foreign substances in the spinning dope solution are easily removed, but the pressure loss in the filter medium increases and the stability of the manufacturing process decreases. Although the tendency described above has been known so far, optimum filtration conditions are different for filter media, and thus any generalizable knowledge has not been obtained for filtration of the spinning dope solution. Accordingly, at the time of changing the filter medium, it has taken a great deal of time and cost to optimize the filtration condition.

[0032] In the method for manufacturing a carbon fiber bundle of the present invention, when the particle retention of the filter medium used for filtering the spinning dope solution is B (µm) and the filter basis weight is D (g/m²), the spinning dope solution is filtered under conditions where a relationship of the filtration speed A (cm/hour), a particle retention B (µm), a filter medium basis weight D (g/m²) satisfies the following equations (1) to (3), then the filtered spinning dope solution is spun to obtain a precursor fiber bundle for carbon fiber.

[0033] Here, the filter basis weight D (g/m²) is a total basis weight of a filter medium main body excluding a mesh layer which may be laminated for the purpose of protecting the filter medium main body. The filter basis weight D can be calculated by measuring the mass of the filter medium cut out into an arbitrary area and dividing this mass by the area.

[0034] As the filter basis weight D is larger, the trapping rate of foreign substances is higher. Conversely, as the filter basis weight D is smaller, foreign substances cannot be easily trapped but tend to slip through. Therefore, when the effect of the filter basis weight D on improvement of the quality of the precursor fiber bundle for carbon fiber and suppression of clogging of the filter was measured while changing the filtration speed A and particle retention B, it was confirmed that there was a minimum filter basis weight that could achieve both improvement of the quality of the precursor fiber bundle for carbon fiber and suppression of clogging of the filter at an arbitrary filtration speed and particle retention (hereinafter described as "minimum filter basis weight"). According to the results of this experiment, the minimum filter basis weight can be expressed using α × β, a product of mutually independent parameters α and β, as shown in the second term on the left side of equation (1). Here, α is defined as a function of the filtration speed A shown in equation (2), and β is defined as a function of the particle retention B shown in equation (3). As the α × β is larger, the minimum filter basis weight is smaller, and as the α × β is smaller, the minimum filter basis weight is larger. As effects of movement of the individual parameters, as the filtration speed A is larger, α is smaller and the minimum filter basis weight is larger. As the filtration speed A is smaller, α is larger and the minimum filter basis weight is smaller. Similarly, as the particle retention B is larger, β is smaller and the minimum filter basis weight is larger. As the particle retention B is smaller, β is larger and the minimum filter basis weight is smaller. Both improvement of the quality of the precursor fiber bundle for carbon fiber and suppression of clogging of the filter can be achieved by performing filtration under conditions satisfying equations (1) to (3). Although this mechanism is not necessarily clarified, it is considered as follows. In other words, as the particle retention is smaller, foreign substances are likely to be caught by a flow path through the filter medium, so that foreign substances can be effectively trapped, whereas the filter is likely to clog. However, it is considered that when the filtration speed is low, deformation and spreading of foreign substances in the filter medium due to pressure drop are suppressed, so that the flow path in the filter medium hardly clogs.

[0035]  In addition, as an example of the method for manufacturing a carbon fiber bundle of the present invention, a filter medium with the particle retention B (µm) satisfying the following equation (4) is preferably used.

[0036] When the particle retention B is 3 or more, suppression of clogging of the filter can be made more effective. Although the reason for this phenomenon is not necessarily clarified, it is considered that, as the value of particle retention B is larger, the filtration pressure tends to decrease, and thus the degree of deformation of foreign substances is so smaller that a filter clogging suppressing effect tends to appear.

[0037] Next, a method for manufacturing a precursor fiber bundle for carbon fiber suitable for obtaining a carbon fiber bundle will be described. In manufacturing a precursor fiber bundle for carbon fiber, it is preferable to obtain a precursor fiber for carbon fiber with a small mean surface roughness on the surface of a single fiber by using a dry-jet wet spinning method. A method for manufacturing a precursor fiber bundle for carbon fiber includes a spinning process for extruding a spinning dope solution from a spinneret into a coagulation bath by a dry-jet wet spinning method and spinning fibers, a water washing process for cleaning the fibers obtained in the spinning process in a water bath, a water bath stretching process for stretching the fibers obtained in the water washing process in a water bath, and a dry-heat treatment process for drying and heat treating the fibers obtained in the water bath stretching process, and may include a steam stretching process for steam stretching the fibers obtained in the dry-heat treatment process may be included, as necessary.

[0038] In the manufacture of a precursor fiber bundle for carbon fiber, the coagulation bath preferably contains a coagulant and a solvent used as a solvent for the spinning dope solution. As the coagulant, a component that does not dissolve a polyacrylonitrile copolymer and is compatible with the solvent used in the spinning dope solution can be used. Specifically, it is preferable to use water as the coagulant.

[0039] In the manufacture of a precursor fiber bundle for carbon fiber, the water bath temperature in the water washing process is preferably 30 to 98°C, and is preferably washed using a water washing bath having a plurality of stages.

[0040] In addition, the stretch ratio in the water bath stretching process is preferably 2 to 6 times.

[0041] After the water bath stretching process, it is preferable to apply an oil agent made of silicone or the like to the fiber bundle for the purpose of preventing adhesion between single fibers. As such a silicone oil agent, it is preferable to use a modified silicone, and it is preferable to use one containing an amino-modified silicone having high heat resistance.

[0042] A known method can be used for the dry-heat treatment process. For example, the drying temperature is 100 to 200°C.

[0043] A precursor fiber bundle for carbon fiber that is more suitably used for the manufacture of a carbon fiber bundle is obtained by further performing the steam stretching process, after the water washing process, water bath stretching process, oil agent application process, and dry-heat treatment process described above. As the steam stretching process, it is preferable to stretch 2 to 6 times in pressurized steam.

[0044] The mean fineness of single fibers contained in the precursor fiber bundle for carbon fiber thus obtained is preferably 0.7 to 1.5 dtex, and more preferably 0.9 to 1.2 dtex. By setting the single-fiber fineness to 0.7 dtex or more, the occurrence of fiber bundle fracture due to the accumulation of single fiber fracture due to contact with rollers and guide parts is suppressed, and the process stability of each of the spinning process, oxidation process, precarbonization process and carbonization process can be maintained. Also, by setting the single-fiber fineness to 1.5 dtex or less, the skin layer ratio in each single fiber after the oxidation process is reduced, the process stability in the subsequent carbonization process and the tensile strength of strands and elastic modulus of strands of the resulting carbon fiber bundle can be improved. In order to adjust the single-fiber fineness of the resulting precursor fiber bundle for carbon fiber, it is only necessary to adjust the extrusion amount of the spinning dope solution in the spinning process for extruding the spinning dope solution from the spinneret and spinning fibers.

[0045] The resulting precursor fiber bundle for carbon fiber is usually continuous fibers. Also, the number of filaments per one fiber bundle is preferably 10,000 to 60,000.

[0046] The method for manufacturing a carbon fiber bundle of the present invention is such that the precursor fiber bundle for carbon fiber is heat-treated in an oxidizing atmosphere until the density reaches 1.32 to 1.35 g/cm³, and then heat-treated at 275°C or more and 295°C or less in an oxidizing atmosphere until the density reaches 1.46 to 1.50 g/cm³. That is, the precursor fiber bundle for carbon fiber is heat-treated until reaching a predetermined density in the former half of the oxidation process, and then heat-treated at a high temperature of 275°C or more and 295°C or less in the latter half of the oxidation process.

[0047] Here, the oxidizing atmosphere is an atmosphere containing 10% by mass or more of a known oxidizing substance such as oxygen and nitrogen dioxide, and an air atmosphere is preferable from the viewpoint of simplicity.

[0048] The density of the oxidized fiber bundle is generally used as an index indicating the progress of the oxidation reaction. When the density is 1.32 g/cm³ or more, the oxidized fiber bundle has a high heat resistant structure, so that it is difficult to be decomposed when heat-treated at a high temperature, and the tensile strength of strands of the resulting carbon fiber bundle is improved. In addition, when the density is 1.35 g/cm³ or less, a long heat treatment time at a high temperature can be secured in the subsequent process, so that the tensile strength of strands of the carbon fiber bundle can be improved. In the oxidation process, in order to enable the process temperature to be switched as described above at the density specified by the oxidized fiber bundle, it is only necessary to collect the fiber bundles during the former half and the latter half of the oxidation process and measure their densities. A method for measuring the density will be described later. For example, when the measured density of the oxidized fiber bundle is lower than specified, the density of the oxidized fiber bundle can be adjusted by raising the temperature or prolonging the oxidation time in the former half of the oxidation process.

[0049] In the oxidation process, first, the precursor fiber bundle for carbon fiber is heat-treated in an oxidizing atmosphere, at preferably 210°C or more and less than 245°C, more preferably 220°C or more and less than 245°C, and further preferably 225°C or more and less than 240°C, thereby obtaining an oxidized fiber bundle with a density of preferably 1.22 to 1.24 g/cm³, and more preferably a density of 1.23 to 1.24 g/cm³. When the density of the oxidized fiber bundle is 1.22 g/cm³ or more, the chemical structure of the single fiber in the oxidation process is stabilized by heat treatment, and the difference between skin-core structure of the single fiber does not deteriorate even when the subsequent heat treatment is performed at a high temperature, so that the tensile strength of strands is often improved. Further, when the density is 1.24 g/cm³ or less, the total amount and time of heat treatment including the subsequent heat treatment is reduced, which is often superior in terms of tensile strength of strands and productivity. Regarding the temperature, a temperature of 210°C or more is preferable because the difference between skin-core structure can be sufficiently suppressed. At a temperature of less than 245°C is preferable because it is an oxidation initial temperature sufficiently low to suppress the difference between skin-core structure regarding the single fiber diameter of the precursor fiber bundle for carbon fiber, the tensile strength of strands is often increased.

[0050] The oxidized fiber bundle is heat-treated until the density reaches 1.22 to 1.24 g/cm³ and then heat-treated in an oxidizing atmosphere to obtain an oxidized fiber bundle with a density of 1.32 to 1.35 g/cm³, and more preferably 1.33 to 1.34 g/cm³. This heat treatment process is performed in an oxidizing atmosphere at preferably 245°C or more and less than 275°C, and more preferably 250°C or more and less than 270°C. When the density is 1.32 g/cm³ or more, the chemical structure of the single fiber in the oxidation process is further stabilized by heat treatment, and the difference between skin-core structure does not deteriorate even when the subsequent heat treatment is performed at a higher temperature, so that the tensile strength of strands is often improved. Further, when the density is 1.35 g/cm³ or less, the total amount and time of heat treatment including the subsequent heat treatment are reduced, and the tensile strength of strands and productivity are superior. When the heat treatment temperature is 245°C or more, the total amount and time of heat treatment are reduced, and the tensile strength of strands and productivity are often superior. When the heat treatment temperature is less than 275°C, even when heat-treating an oxidized fiber bundle with a density of 1.22 to 1.24 g/cm³, the difference between skin-core structure can be suppressed, and high tensile strength of strands is often exhibited.

[0051] Subsequently, the obtained oxidized fiber bundle is heat-treated in an oxidizing atmosphere at a temperature of 275°C or more to 295°C or less, and preferably 280°C or more to 290°C or less to obtain an oxidized fiber bundle with a density of 1.46 to 1.50 g/cm³. When the heat treatment temperature is 275°C or more, the amount of heat applied when increasing the density can be reduced, whereby the tensile strength of strands is improved. When the heat treatment temperature is 295°C or less, it is possible to proceed the oxidation reaction without decomposing the structure of the single fiber, and maintain the tensile strength of strands. In order to measure the heat treatment temperature, it is only necessary to insert a thermometer such as a thermocouple into a heat treatment oven in the oxidation process to measure oven temperature. In a case where there are temperature unevenness and temperature distribution when measuring the oven temperature in several points, the simple average temperature is calculated.

[0052] In the present invention, the final density of the oxidized fiber bundle is 1.46 to 1.50 g/cm³, preferably 1.46 to 1.49 g/cm³, and further preferably 1.47 to 1.49 g/cm³. Since the density of the oxidized fiber bundle correlates with the carbonization yield, the higher density is better, from the viewpoint of reducing manufacturing energy. When the density is 1.46 g/cm³ or more, the carbonization yield can be sufficiently increased. When the density is 1.50 g/cm³ or less, the effect of increasing the carbonization yield is not saturated, which is effective from the viewpoint of productivity. In order to complete the heat treatment at the specified density, it is only necessary to adjust the oxidation temperature and time.

[0053] In the process of heat treatment at 275°C or more and 295°C or less in an oxidizing atmosphere until the density of the oxidized fiber bundle reached 1.46 to 1.50 g/cm³, the tension applied to the oxidized fiber bundle (oxidation tension) is preferably 1.6 to 4.0 mN/dtex, more preferably 2.5 to 4.0 mN/dtex, and further preferably 3.0 to 4.0 mN/dtex. The oxidation tension is represented by a value obtained by dividing the tension (mN) measured on an exit side of the oxidation oven by the fineness (dtex) of the precursor fiber bundle for carbon fiber in complete dryness. When the tension is 1.6 mN/dtex or more, the orientation of the carbon fiber bundle is sufficiently increased, and the tensile strength of strands is often improved. When the tension is 4.0 mN/dtex or less, quality deterioration due to fuzz tends to be small

[0054]  Generally, when the density of the oxidized fiber bundle is increased in order to obtain a high carbonization yield, the tensile strength of strands of the carbon fiber bundle tends to decrease. In the method for manufacturing a carbon fiber bundle of the present invention, even when the density of the oxidized fiber bundle is increased by performing high-temperature heat treatment in the latter half at an appropriate temperature profile in the oxidation process, the difference between skin-core structure of the single fiber is greatly suppressed, and the structure is stabilized, so that both high carbonization yield and high tensile strength of strands can be achieved.

[0055] Except for the oxidation process, a known method for manufacturing a carbon fiber bundle may be basically followed. However, in the method for manufacturing a carbon fiber bundle of the present invention, it is preferable to perform a pre-carbonization process, following the spinning process and the oxidation process. In the pre-carbonization process, it is preferable to obtain a carbonized fiber bundle by heat-treating the oxidized fiber obtained in the oxidation process in an inert atmosphere at a maximum temperature of 500 to 1000°C until the density reaches 1.5 to 1.8 g/cm³.

[0056]  A carbonization process is performed, following the pre-carbonization. In the carbonization process, it is preferable to obtain a carbon fiber bundle by heat-treating the pre-carbonized fiber bundle in an inert atmosphere at a maximum temperature of 1200 to 1800°C, and preferably 1200 to 1600°C. When the maximum temperature is 1200°C or more, the nitrogen content in the carbon fiber bundle is reduced, and the tensile strength of strands is stably exhibited. When the maximum temperature is 1800°C or less, a satisfactory carbonization yield can be obtained.

[0057] The carbon fiber bundle obtained as described above is preferably subjected to an oxidation treatment so that an oxygen containing functional group is introduced, in order to improve adhesion to a matrix resin. As the oxidation treatment method, gas phase oxidation, liquid phase oxidation, liquid phase electrolytic oxidation and the like are used. From the viewpoint that high productivity and uniform treatment can be achieved, liquid phase electrolytic oxidation is preferably used. The method of liquid phase electrolytic oxidation is not particularly specified, and may be performed by a known method.

[0058] After such an electrochemical treatment, a sizing treatment can also be performed to impart convergency to the obtained carbon fiber bundle. As the sizing agent, a sizing agent having good compatibility with the matrix resin can be appropriately selected according to the type of the matrix resin used for a composite material.

[0059] The measuring methods of various physical property values described in this specification are as follows.

<Tensile Strength of Strands and Elastic Modulus of Strands of Carbon Fiber Bundle>

[0060] The tensile strength of strands and elastic modulus of strands of the carbon fiber bundle are determined in accordance with a resin-impregnated strand test method of JIS-R-7608 (2004), according to the following procedure. Ten resin-impregnated strands of the carbon fiber bundle are measured, and the average value thereof is defined as the tensile strength of strands. Strain is evaluated using an extensometer. The strain was evaluated at a strain range of 0.1 to 0.6%. As a resin formulation, "CELLOXIDE (registered trademark)" 2021P (manufactured by Daicel Chemical Industries, Ltd.)/boron trifluoride monoethylamine (manufactured by Tokyo Chemical Industry Co., Ltd.)/acetone = 100/3/4 (parts by mass) are used. As curing conditions, atmospheric pressure, a temperature of 125°C, and a time of 30 minutes are used.

<Density Measurement>

[0061] 1.0 to 3.0 g of the oxidized fiber bundle is collected and completely dried at 120°C for 2 hours. Next, after measuring an absolute dry mass A (g), the oxidized fiber bundle is impregnated with ethanol and sufficiently defoamed, then a fiber mass B (g) in the ethanol solvent bath is measured, and a density is determined by density = (A × ρ)/(A - B). ρ is a specific gravity of ethanol at the measurement temperature.

<Skin Layer Ratio of Single Fiber of Carbon Fiber>

[0062] A carbon fiber bundle to be measured is embedded in a resin, a cross section perpendicular to the fiber axis direction is polished, and the cross section is observed using a 100 times objective lens of an optical microscope at a total magnification of 1000. The blackened thickness of the outer peripheral portion is measured from the cross-sectional microscopic image of the polished surface. Analysis is performed using image analysis software Image J. First, in a single fiber cross-sectional image, black and white area division is performed by binarization. For the luminance distribution in the single fiber cross section, the average value of the distribution is set as a threshold value, and binarization is performed. In the obtained binarized image, the shortest distance from a point on the surface layer to a lined region from black to white is measured in the fiber diameter direction. This is measured for five points in the circumference of the same single fiber, and the average value is calculated as the blackened thickness at that level. Further, the skin layer ratio is calculated from the area ratio (%) of the blackened thickness portion with respect to the entire cross section perpendicular to the fiber axis direction of the single fiber of the carbon fiber. The same evaluation is performed on 30 single fibers in the carbon fiber bundle, and the average value thereof is used.

<Average Single Fiber Diameter of Carbon Fiber Bundle>

[0063] A mass A_f (g/m) and a density B_f (g/cm³) per unit length are determined for a carbon fiber bundle composed of a large number of carbon filaments to be measured. The number of filaments of the carbon fiber bundle to be measured is defined as C_f, and the average single fiber diameter (µm) of the carbon fibers is calculated by the following equation.

[0064]  <Knot Strength of Carbon Fiber Bundle>

[0065] A grip part having a length of 25 mm is attached to both ends of a carbon fiber bundle with a length of 150 mm to prepare a test specimen. In the preparation of the test specimen, a load of 9.0 × 10^-5 N/dtex is applied to the carbon fiber bundle for alignment. One knot is made at the midpoint of the test specimen, and the test specimen is subjected to a fiber bundle tensile test at a crosshead speed at tension of 100 mm/min. A total of 12 fiber bundles are subjected to the measurement. The average value of 10 fiber bundles excluding the maximum value and the minimum value is used as the measured value. As the knot strength, a value obtained by dividing the maximum load value obtained in the fiber bundle tensile test by the average cross-sectional area value of the carbon fiber bundles is used.

<Probability That Flaw with Size of 50 nm or More Exists>

[0066] A single fiber tensile test of the single fibers of the carbon fibers is performed in accordance with JIS R7606 (2000), and a sample of the single fibers of the carbon fibers after fracture including a fracture surface (hereinafter simply referred to as "fracture surface") is collected. The number of single fibers to be tested is one set of 50 fibers, and when it is not possible to collect 30 or more pairs of fracture surfaces on both sides, another set of 50 fibers is subjected to a single fiber tensile test to collect 30 or more pairs of fracture surfaces on both sides. The strain rate during the tensile test is set to 0.4 mm/min.

[0067] From the pairs of fracture surfaces collected as described above, 30 pairs are randomly selected and observed with a scanning electron microscope (SEM). Before the observation, a vapor deposition treatment for applying conductivity is not performed, and the observation is performed at an acceleration voltage of 1 keV and at a magnification of 25,000 to 50,000. In addition, in order to make it easy to determine the presence or absence of minute flaws, a stage is rotated such that the fracture origin faces the front side, and the stage is tilted by 30° to observe the fracture origin from the oblique upside (see Figs. 1 to 4).

[0068] Because traces of the fracture radially progressing from the fracture origin (i) remained as radial streaks on the original fracture surface caused by tensile fracture of the carbon fiber, a portion on which the streaks present on an SEM observation image converged to one point when traced is identified as a fracture origin (i). When the streaks cannot be recognized or when the streaks can be recognized, but stain is adhered near the fracture origin (i) so that the streaks are hardly observed on at least one side of the fracture surfaces on both sides, the pair of such fracture surfaces is excluded from evaluation. The fracture surface reduced by the exclusion is replenished as appropriate so that 30 pairs of fracture surfaces will eventually be observed.

[0069] Once the fracture origin (i) can be identified, it is examined whether there are any morphological features. There are various types of morphological features such as dents, attached substances, traces of the fiber surface being partially peeled, damages and adhesion marks. The morphological features to be fracture origins that can be observed by SEM are collectively referred to as "flaws". Lengths measured along the circumferential direction of the fiber, that is, those with a size of 50 nm or more, are uniformly classified as a "fracture surface in which a flaw with a size of 50 nm or more exists" in the present invention regardless of differences in appearance. When it is performed on the fracture surfaces on both sides and either one is classified as the "fracture surface in which a flaw with a size of 50 nm or more exists", the pair is taken as having the "fracture surface in which a flaw with a size of 50 nm or more exists". This classification is performed on all 30 pairs of fracture surfaces observed with SEM, and the total number of "fracture surfaces in which a flaw with a size of 50 nm or more exists" is divided by 30 which is the total number of the pairs of fracture surfaces observed by SEM and multiplied by 100 to calculate a "probability (%) that a flaw with a size of 50 nm or more exists".

[0070] Here, the single fiber tensile test was performed by TENSILON "RTC-1210A" manufactured by A&D Company, Limited, with a gauge length of 10 mm, using a commercially available cyanoacrylate instant adhesive for fixing the carbon fiber to a test piece mount, and using a special test jig designed to perform in water. Further, a scanning electron microscope (SEM) "S-4800" manufactured by Hitachi High-Technologies Corporation was used for observing the collected fracture surfaces.

<Mean Surface Roughness>

[0071] Using ten single fibers of the carbon fibers to be evaluated that are placed on a sample stage and fixed with an epoxy resin as samples, evaluation is performed using an atomic force microscope (in Examples, NanoScope V Dimension Icon, manufactured by Bruker AXS). In Examples, a three-dimensional surface shape image is obtained under the following conditions.

Probe: silicon cantilever (OMCL-AC160TS-W2 manufactured by Olympus)

Measurement mode: tapping mode

Scanning speed: 1.0 Hz

Scanning range: 600 nm × 600 nm

Resolution: 512 pixels × 512 pixels

Measurement environment: room temperature, in air.

[0072] For a single fiber, a three-dimensional surface shape image is measured under the above conditions, and the obtained measurement image is subjected to image processing using, "flat treatment" for removing undulation of data derived from the device using the attached software (NanoScope Analysis), taking a curvature of a fiber cross section into account, "median 8 treatment" which is a filter treatment that replaces a central value of a matrix from a median value of Z data in the 3 × 3 matrix, and "three-dimensional tilt correction" for carrying out fitting of a cubic curved surface by a least square method from all image data and correcting in-plane tilt, then surface roughness analysis is performed with the attached software to calculate mean surface roughness. Here, the mean surface roughness (Ra) is a three-dimensional extension of a centerline roughness Ra defined in JIS B0601 (2001) so that it can be applied to surface measurement and is defined as a mean value of absolute values of deviations from a reference surface to a designated surface. As to measurement, 10 different single fibers are randomly sampled, and the measurement is performed once for each single fiber, 10 times in total, and the average value thereof is taken as the measured value.

<Number of Fuzzes of Carbon Fiber Bundle>

[0073] The quality of the carbon fiber bundle, which affects productivity during manufacture of the composite material, is evaluated by a method of directly counting the number of fuzzes by the following method. By visually observing the running carbon fiber bundle at a running speed of 1.5 m/min and a stretch ratio of 1 time, the number of fractures single fibers protruding 5 mm or more from the surface of the carbon fiber bundle is counted at a length of the carbon fiber bundle of 20 m to evaluate the number of fuzzes per 1 m (fuzzes/m).

[Examples]

(Example 1)

[0074] A copolymer composed of 99% by mass of acrylonitrile and 1% by mass of itaconic acid was polymerized by solution polymerization using dimethyl sulfoxide as a solvent to produce a polyacrylonitrile copolymer to obtain a spinning dope solution. The spinning dope solution was allowed to flow into a filter device and filtered. The filter medium used was a sintered metal filter having a particle retention B of 1 µm, a filter medium thickness C of 800 µm, and a filter basis weight D of 2500 g/m², and filtration was performed under a filtration condition with a filtration speed A of 3 cm/hour. Fibers were spun by a dry-jet wet spinning method in which the filtered spinning dope solution was once extruded through a spinneret into the air and introduced into a coagulation bath composed of an aqueous solution of 35% dimethyl sulfoxide controlled at 3°C. The spun fiber bundle was washed with water at 30 to 98°C, and 3.5 times water bath stretching was performed at that time. Subsequently, an amino-modified silicone-based silicone oil agent was applied to the fiber bundle after the water bath stretching in the water bath and dried using a roller heated to a temperature of 160°C to obtain a fiber bundle with a number of single fibers of 12,000. The fiber bundle was stretched 3.7 times in pressurized steam to make the total stretch ratio of the yarn 13 times. Then, the fiber bundle was subjected to entangling treatment by air having a fluid extrusion pressure of 0.35 MPa with a tension of 2 mN/dtex being applied to the fiber bundle to obtain a precursor fiber bundle for carbon fiber with a single-fiber fineness of 1.1 dtex and a number of single fibers of 12,000. Next, using the oxidizing conditions described in Condition 1 in Table 1, the precursor fiber bundle for carbon fiber was heat-treated in an oven in an air atmosphere at a stretch ratio of 1.0 to obtain an oxidized fiber bundle.

[0075] The obtained oxidized fiber bundle was subjected to a pre-carbonization treatment at a stretch ratio of 0.95 in a nitrogen atmosphere at a temperature of 300 to 800°C to obtain a pre-carbonized fiber bundle. The obtained pre-carbonized fiber bundle was subjected to a carbonization treatment at a maximum temperature of 1350°C in a nitrogen atmosphere. The obtained carbon fiber bundle was subjected to surface treatment and sizing agent coating treatment to obtain a final carbon fiber bundle. The number of fuzzes of the carbon fiber bundle at this time was less than 0.1/m, and almost no fuzz was confirmed and the quality was good.

[0076] Table 2 shows the tensile strength of strands, elastic modulus of strands, skin layer ratio of the single fibers of the carbon fibers, and average single fiber diameter of the obtained carbon fiber bundle.

[Table 1]

	Oxidized fiber bundle density			Oxidation temperature			Oxidation tension
	First oven exit	Second oven exit	Final oven exit	First oven	Second oven	Final oven	Final oven
	g/cm3			°C			mN/dtex
Condition 1	1.23	1.32	1.48	235	260	285	1.5
Condition 2	1.23	1.32	1.48	225	260	285	1.5
Condition 3	1.23	1.32	1.48	245	260	285	1.2

[Table 2-1]

	Filtration conditions
	Filtration speed A	Particle retention B	Filter medium thickness C	Filter basis weight D	α	β	D-600 / (α×β)
	cm/h	µm	µm	g/m3	-	-	-
Example 1	3	1	800	2500	0.98	0.44	1120
Example 2	3	9	3200	6400	0.98	0.11	947
Example 3	6	1	800	2500	0.73	0.44	646
Example 4	6	1	800	2500	0.73	0.44	646
Example 5	6	1	800	2500	0.73	0.44	646
Comparative Example 1	3	9	1600	3200	0.98	0.11	-2253
Comparative Example 2	6	9	1600	3200	0.73	0.11	-4125
Comparative Example 3	6	9	3200	6400	0.73	0.11	-925
Comparative Example 4	8	1	800	2500	0.27	0.44	-2539
Comparative Example 5	12	1	800	2500	0.01	0.44	-199979
Example 6	3	1	800	2500	0.98	0.44	1120
Example 7	3	1	800	2500	0.98	0.44	1120

[Table 2-2]

	Carbon fiber bundle
	Tensile strength of strands	Elastic modulus of strands	Probability that flaw with size of 50 nm or more exists	Skin layer ratio of single fiber of carbon fiber	Average single fiber diameter d	Knot strength K	- 88d+1390	Mean surface roughness Ra
	GPa	GPa	%	%	µm	MPa	MPa	nm
Example 1	5.9	256	17	91	7.5	736	730	-
Example 2	5.8	257	22	91	7.5	752	730	-
Example 3	5.8	253	29	91	7.5	741	730	1.5
Example 4	6.0	250	29	91	7.3	772	748	-
Example 5	6.2	263	29	91	7.0	785	774	1.4
Comparative Example 1	5.7	252	52	91	7.5	722	730	-
Comparative Example 2	5.6	257	57	91	7.4	778	739	2.1
Comparative Example 3	5.6	247	38	91	7.5	698	730	-
Comparative Example 4	5.4	261	37	91	7.5	838	730	-
Comparative Example 5	5.2	262	81	91	7.5	867	730	-
Example 6	5.8	261	17	97	7.5	731	730	-
Example 7	5.8	245	17	85	7.5	732	730	-

(Example 2)

[0077] A precursor fiber bundle for carbon fiber and a carbon fiber bundle were obtained in the same manner as in Example 1, except that the filter medium was changed to a sintered metal filter having a particle retention B of 9 µm, a filter medium thickness C of 3200 µm, and a filter basis weight D of 6400 g/m².

(Example 3)

[0078] A precursor fiber bundle for carbon fiber and a carbon fiber bundle were obtained in the same manner as in Example 1 except that the filtration speed A was changed to 6 cm/hour under the filtration conditions.

(Examples 4 and 5)

[0079] A precursor fiber for carbon fiber and a carbon fiber bundle were obtained in the same manner as in Example 3, except that the stretch ratio during pre-carbonization was 1.05 times in Example 4 and 1.10 times in Example 5.

(Comparative Example 1)

[0080] A precursor fiber bundle for carbon fiber and a carbon fiber bundle were obtained in the same manner as in Example 2, except that the filter medium was changed to a sintered metal filter having a filter medium thickness C of 1600 µm and a filter basis weight D of 3200 g/m². The number of fuzzes of the carbon fiber bundle was 0.2/m, and the quality deteriorated.

(Comparative Example 2)

[0081] A precursor fiber bundle for carbon fiber and a carbon fiber bundle were obtained in the same manner as in Comparative Example 1 except that the filtration speed A was changed to 6 cm/hour under the filtration conditions.

(Comparative Example 3)

[0082] A precursor fiber bundle for carbon fiber and a carbon fiber bundle were obtained in the same manner as in Example 2 except that the filtration speed A was changed to 6 cm/hour under the filtration conditions.

(Comparative Example 4)

[0083] A precursor fiber bundle for carbon fiber and a carbon fiber bundle were obtained in the same manner as in Example 3 except that the filtration speed A was changed to 8 cm/hour under the filtration conditions.

(Comparative Example 5)

[0084] A precursor fiber bundle for carbon fiber and a carbon fiber bundle were obtained in the same manner as in Example 3 except that the filtration speed A was changed to 12 cm/hour under the filtration conditions.

(Example 6)

[0085] A carbon fiber bundle was obtained in the same manner as in Example 1 except that Condition 2 in Table 1 was used as the oxidizing condition. The skin layer ratio of the carbon fibers was 97%, and the tensile strength of strands decreased as compared to that in Example 1.

(Example 7)

[0086] A carbon fiber bundle was obtained in the same manner as in Example 1 except that Condition 3 in Table 1 was used as the oxidizing condition. The skin layer ratio of the carbon fibers was 85%, and the tensile strength of strands decreased as compared to that in Example 1.

INDUSTRIAL APPLICABILITY

[0087] The present invention can obtain an oxidized fiber bundle having a specific density by heat-treating at an appropriate temperature profile in the oxidation process, whereby flaws governing the tensile strength of strands and the knot strength are controlled to be very small, and thus can manufacture a carbon fiber bundle that exhibits the tensile strength of strands and the elastic modulus of strands in a well-balanced manner and also exhibits high knot strength without impairing productivity. Moreover, according to the carbon fiber bundle of the present invention, it becomes a carbon fiber bundle which satisfies the productivity at the time of manufacturing a composite material. The carbon fiber bundle to be obtained by the present invention is suitably used for general industrial uses such as aircraft, automobile and ship members, sports uses such as golf shafts and fishing rods, and pressure vessels, taking advantage of such characteristics.

DESCRIPTION OF REFERENCE SIGNS

[0088] (i): Fracture origin

Claims

1. A method for manufacturing a carbon fiber bundle, the method comprising filtering a spinning dope solution in which a polyacrylonitrile copolymer is dissolved in a solvent, using a filter medium having a particle retention B (µm) and a filter basis weight D (g/m²), under conditions where a filtration speed A (cm/hour) satisfies following equations (1) to (3), then spinning the filtered spinning dope solution to obtain a precursor fiber bundle for carbon fiber,





heat-treating the obtained precursor fiber bundle for carbon fiber in an oxidizing atmosphere until a density reaches 1.32 to 1.35 g/cm³, then heat-treating at 275°C or more and 295°C or less in an oxidizing atmosphere until a density reaches 1.46 to 1.50 g/cm³ to obtain an oxidized fiber bundle, and then heat-treating the oxidized fiber bundle at 1200 to 1800°C in an inert atmosphere.

2. The method according to claim 1, wherein a tension of the oxidized fiber bundle when heat-treated at 275°C or more and 295°C or less in an oxidizing atmosphere until the density reaches 1.46 to 1.50 g/cm³ is 1.6 to 4.0 mN/dtex.

3. The method according to claim 1 or 2, further comprising heat-treating the precursor fiber bundle for carbon fiber at 210°C or more and less than 245°C in an oxidizing atmosphere until the density reaches 1.22 to 1.24 g/cm³, and then subjecting to the heat treatment process in an oxidizing atmosphere until the density reaches 1.32 to 1.35 g/cm³, wherein the heat treatment process performed until the density reaches 1.32 to 1.35 g/cm³ is performed at a temperature of 245°C or more and less than 275°C.

4. A carbon fiber bundle having an elastic modulus of strands of 240 to 280 GPa, a tensile strength of strands of 5.8 GPa or more, a knot strength K [MPa] of -88d + 1390 ≤ K (d: average single fiber diameter [µm]), and an average single fiber diameter of 6.5 to 8.0 µm, wherein a probability that a flaw with a size of 50 nm or more exists on a fracture surface, which is collected when a single fiber tensile test is performed with a gauge length of 10 mm, is 35% or less.

5. The carbon fiber bundle according to claim 4, having a knot strength K of 770 MPa or more.

6. The carbon fiber bundle according to claim 4 or 5, having a mean surface roughness Ra of 1.0 to 1.8 nm.

7. The carbon fiber bundle according to any one of claims 4 to 6, wherein a skin layer ratio of the single fibers of the carbon fibers is 90% or more by area.

Drawing