Carbon fiber bundle and method for manufacturing the same

Abstract

[Problem] To provide: a carbon fiber bundle that, as a result of specific fluff inside the carbon fiber bundle being suppressed, can suppress catching caused by ring-shaped fluff produced when the carbon fiber bundle is unwound for higher-order processing; and a production method for the carbon fiber bundle. [Solution] A carbon fiber bundle that has an average single fiber diameter B of 6.9–11.0 μm and a resin-impregnated strand tensile elastic modulus E of 230–310 GPa, there being no more than 40 pieces of fluff per m inside the carbon fiber bundle, and the fraction of the fluff inside the carbon fiber bundle that is fluff that has a two-layer cross-section being 1%–25%. It is suitable to obtain the carbon fiber bundle by heat treating a polyacrylonitrile precursor fiber bundle that has a single-fiber fineness of 0.9–2.2 dtex in a 200°C–300°C oxidizing atmosphere to obtain a flame-resistant fiber bundle and then heat treating the flame-resistant fiber bundle at 1,200°C–1,600°C in an inert atmosphere, the heat treatment of the polyacrylonitrile precursor fiber bundle being performed such that the heat generation rate Q on the left side of expression (3) (in which q (J / g / s) is the single-fiber heat generation rate, N (strands) is the filament count, d (dtex) is the single-fiber fineness of the flame-resistant fiber bundle, and W (mm) is the thread width) is 150–500 J / m2 / s until the density is 1.22–1.24 g / cm3 and then the heat treatment being continued while the tension of the flame-resistant fiber bundle applies 1.6–4.0 mN / dtex until the density is 1.38–1.50 g / cm3. (3) Q=q×N×d / W / 10.

Description

[Technical Field]

[0001] The present invention provides a carbon fiber bundle and a method for manufacturing the same, which have high load-bearing capacity per single fiber and excellent abrasion resistance, and which can suppress specific fuzz inherent in the carbon fiber bundle even when the single fiber fineness is high, thereby suppressing entanglement caused by ring-shaped fuzz that occurs when the carbon fiber bundle is unwound for higher-order processing. [Background technology]

[0002] Composite materials using carbon fiber bundles are used in aerospace applications, as well as sports applications such as bicycles and golf clubs. Recently, their use has expanded to industrial applications such as automotive components and pressure vessels. In industrial applications, where reducing production costs is crucial, it is important to improve processability, such as improving the abrasion resistance of these components with rollers during molding and suppressing fluffing (single fiber breakage) when carbon fiber bundles are unwound or run on rollers. In particular, suppressing the ring-shaped fluff generated when carbon fiber bundles are unwound is important because it can cause the surrounding carbon fiber bundles to become entangled with the rollers.

[0003] Generally, polyacrylonitrile-based carbon fiber bundles are manufactured by a flame-retardant process in which a polyacrylonitrile-based precursor fiber bundle is oxidized in air at 200-300°C, a pre-carbonization process in which it is heated in an inert atmosphere at 500-1,200°C, and a carbonization process in which it is heated in an inert atmosphere at 1,200-3,000°C. To improve abrasion resistance by increasing the load-bearing capacity per single fiber of the carbon fiber bundle, it is effective to increase the weight per single fiber, i.e., the fineness of the single fiber. To achieve this, it is effective to increase the amount of heat treatment in the flame-retardant process to increase the yield of carbon fiber bundles or to increase the fineness of the single fiber of the polyacrylonitrile-based precursor fiber bundle.

[0004] To date, methods for manufacturing carbon fiber bundles that suppress fluffing during production have been proposed (Patent Documents 1-4).

[0005] Patent Document 1 proposes that by increasing the tension of the carbon fiber bundle during the carbonization process while maintaining the carbonization temperature at 1,000 to 1,500°C, both the tensile modulus of the resin-impregnated strand (hereinafter sometimes abbreviated as strand tensile modulus E) and the compressive strength can be improved, thereby reducing fluffing when the carbon fiber bundle is rubbed against a roller. Patent Document 2 proposes that by controlling the heat treatment temperature of the flame-retardant process according to the density of the flame-retardant fiber bundle during the flame-retardant process, the double structure of the carbon fiber bundle can be suppressed, and furthermore, fluffing can be reduced due to the large single fiber fineness and high knot strength. Patent Document 3 proposes that by controlling the flame-retardant time during the flame-retardant process to satisfy an appropriate flame-retardant structure, the double structure of the flame-retardant fiber bundle can be suppressed. Patent Document 4 proposes that by using hydroxyalkyl methacrylate as a copolymer component to control the amount of heat generated during the flame-retardant process, a carbon fiber bundle with excellent handling and processability can be obtained, even with a large single fiber fineness, because of the high knot strength. [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2005-344254 [Patent Document 2] Japanese Patent Publication No. 2017-66580 [Patent Document 3] Japanese Patent Publication No. 2018-178344 [Patent Document 4] International Publication No. 2013 / 157613 brochure [Overview of the project] [Problems that the invention aims to solve]

[0007] However, the underlying technology has the following challenges.

[0008] In Patent Document 1, the temperature and time of the flame-retardant treatment process were not controlled, and the single fiber fineness of the carbon fiber bundle was small, resulting in low load-bearing capacity per single fiber and insufficient abrasion resistance. Furthermore, while the average physical properties were increased to improve the average resistance to abrasion against rollers, there was a problem in that the ring-shaped fuzzing caused by inherent fuzz that occurs only when the carbon fiber bundle is unwound was not suppressed. In Patent Documents 2 and 3, although the average state was such that fuzzing was less likely to occur, the high single fiber fineness prevented control of the heat generation rate and heat removal rate in the flame-retardant treatment process, resulting in temperature variations within the flame-retardant fiber bundle, which prevented the suppression of specific fuzz inherent in the carbon fiber bundle. In addition, there was a problem in that the ring-shaped fuzzing that occurs when the carbon fiber bundle is unwound for higher-order processing was not suppressed due to the influence of inherent fuzz. In Patent Document 4, although the heat generation rate in the flame-retardant process was controlled, the high fineness of the single fibers of the carbon fiber bundles, coupled with the lack of control over the heat removal rate, resulted in temperature variations within the flame-retardant fiber bundles. This made it impossible to suppress specific fluffing inherent in the carbon fiber bundles, and consequently, it was impossible to suppress the entanglement caused by ring-shaped fluffing that occurred when the carbon fiber bundles were unwound for higher-level processing.

[0009] As described above, it has been proposed that fuzzing generated when carbon fiber bundles run on rollers during the manufacturing process and the process of using carbon fiber bundles can be suppressed by increasing the knot strength, strand tensile modulus E, and compressive strength, as described in Patent Documents 1-2 and 4. It has also been proposed that the double structure can be suppressed by controlling the temperature, time, and heat generation rate of the flame-retardant process, as described in Patent Documents 2-4. Thus, although there have been methods that suppress the double structure on average, they still contain variations, and it was not recognized that some of the fuzzing would have an adverse effect. In other words, none of the inventions considered heat removal within the flame-retardant fiber bundle during the flame-retardant process, so they could not suppress the specific fuzzing inherent in the carbon fiber bundle caused by temperature variations when the single fiber fineness is large, and they could not suppress the entanglement caused by ring-shaped fuzzing that occurs when the carbon fiber bundle is unwound for higher-order processing.

[0010] The present invention aims to provide a carbon fiber bundle and a method for manufacturing the same, which have high load-bearing capacity per single fiber, excellent abrasion resistance, and can suppress specific fuzz inherent in the carbon fiber bundle even with high single fiber fineness, thereby suppressing entanglement caused by ring-shaped fuzz that occurs when the carbon fiber bundle is unwound for higher-order processing. [Means for solving the problem]

[0011] To achieve this objective, the present invention has the following configuration.

[0012] In other words, the carbon fiber bundle of the present invention is characterized by having an average single fiber diameter B of 6.9 to 11.0 μm, a resin-impregnated strand tensile modulus E of 230 to 310 GPa, having 40 or fewer fluffs per meter within the carbon fiber bundle, and having 1 to 25% of the fluffs within the carbon fiber bundle having a double-layered cross-section.

[0013] Furthermore, in the method for producing carbon fiber bundles of the present invention, in the step of heat-treating a polyacrylonitrile-based precursor fiber bundle with a single fiber fineness of 0.9 to 2.2 dtex in an oxidizing atmosphere at 200 to 300°C, when the heat generation rate of the single fiber is q (J / g / s), the number of filaments N is N (threads), the single fiber fineness of the flame-resistant fiber bundle is d (dtex), and the yarn width is W (mm), the density is 1.22 to 1.24 g / cm³. 3 The heat generation rate Q, calculated using equation (3), is 150-500 J / m³ until it reaches this point. 2 After heat treatment to achieve a density of 1.38-1.50 g / cm³, the density is 1.38-1.50 g / cm³. 3 The method is characterized by first heat-treating the fiber bundles to a tension of 1.6 to 4.0 mN / dtex until flame-resistant fiber bundles are obtained, and then heat-treating the flame-resistant fiber bundles in an inert atmosphere at 1,200 to 1,600°C to obtain carbon fiber bundles. Q = q × N × d / W / 10 ... (3) [Effects of the Invention]

[0014] According to the present invention, it is possible to provide a carbon fiber bundle and a method for manufacturing the same, which have high load-bearing capacity per single fiber and excellent abrasion resistance, and which can suppress specific fuzz inherent in the carbon fiber bundle even when the single fiber fineness is high, thereby suppressing entanglement caused by ring-shaped fuzz that occurs when the carbon fiber bundle is unwound for higher-order processing. [Brief explanation of the drawing]

[0015] [Figure 1] This is an example of a photograph showing the structure of a carbon fiber bundle, specifically a cross-section of the fluff containing both an inner and outer layer. [Figure 2] This is an example of a photograph showing a cross-section of a fiber fluff within a carbon fiber bundle, where a hole is present in the center of the fluff's cross-section. [Figure 3] This is an example of a photograph showing a cross-section of a fiber bundle containing fuzz, specifically a fuzz that appears to have fractured due to bending. [Figure 4] This is an example of a photograph showing a deformed cross-section of the fluff within a carbon fiber bundle. [Figure 5] This is an example of a photograph showing a cross-section of a carbon fiber bundle where the fluff within the fiber is perpendicular and splits from the center. [Modes for carrying out the invention]

[0016] The inventors have arrived at the present invention by finding that, in manufacturing a carbon fiber bundle that has a high load-bearing capacity per single fiber, excellent abrasion resistance, and can suppress specific fluffing inherent in the carbon fiber bundle even with a high single fiber fineness, thereby suppressing entanglement caused by ring-shaped fluffing that occurs when the carbon fiber bundle is unwound for higher-order processing, it is possible to sufficiently secure the amount of heat removed relative to the total heat generated in the flame-retardant fiber bundle, thereby reducing temperature variations within the flame-retardant fiber bundle even with a high single fiber fineness.

[0017] First, let's describe the carbon fiber bundle of the present invention.

[0018] The carbon fiber bundle of the present invention has an average single fiber diameter B of 6.9 to 11.0 μm, preferably 7.0 to 10.0 μm, and more preferably 7.1 to 9.0 μm. If the average single fiber diameter B is 6.9 μm or more, fuzzing due to friction can be suppressed, thus suppressing the fuzzing that occurs when the carbon fiber bundle is unwound. If the average single fiber diameter B is 11.0 μm or less, the double structure of the carbon fiber bundle can be suppressed, thus suppressing entanglement due to ring-shaped fuzz that occurs when the carbon fiber bundle is unwound for higher-order processing. The average single fiber diameter B can be calculated from the mass and density per unit length of the carbon fiber bundle and the number of filaments N. Such an average single fiber diameter B can be achieved by controlling the discharge rate in the manufacturing process of the polyacrylonitrile-based precursor fiber bundle, the stretching ratio in each process, and the specific gravity of the flame-resistant fiber bundle.

[0019] The carbon fiber bundle of the present invention has a strand tensile modulus E of 230 to 310 GPa in a resin-impregnated strand tensile test, preferably 245 to 300 GPa, and more preferably 250 to 290 GPa. If the strand tensile modulus E is 230 GPa or higher, a satisfactory modulus of elasticity can be obtained when it is generally used for elastic modulus reinforcement. If the strand tensile modulus E is 310 GPa or lower, fuzzing due to abrasion can be suppressed, and thus, entanglement due to ring-shaped fuzz that occurs when the carbon fiber bundle is unwound for higher-order processing can be suppressed. The strand tensile modulus E can be determined by the method of 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 strand tensile modulus E of the carbon fiber bundle can be controlled mainly by applying tension to the fiber bundle during one of the heat treatment processes in the manufacturing process of the carbon fiber bundle, by improving the double structure, or by changing the carbonization temperature.

[0020] The carbon fiber bundle of the present invention preferably has a crystallite size Lc of 1.5 to 2.5 nm, more preferably 1.6 to 2.3 nm, and even more preferably 1.7 to 2.2 nm. A crystallite size Lc of 1.5 nm or more is preferable because it suppresses the formation of ring-shaped fluff when the carbon fiber bundle is unwound. A crystallite size Lc of 2.5 nm or less is preferable because it eliminates the need to raise the maximum temperature of the carbonization process unnecessarily, resulting in excellent abrasion resistance and suppression of ring-shaped fluff when the carbon fiber bundle is unwound. The crystallite size Lc can be measured by a known method using a wide-angle X-ray diffractometer, and the Scherrer constant in Scherrer's equation described later is 1. Such a crystallite size Lc can be controlled by changing the carbonization temperature.

[0021] The carbon fiber bundle of the present invention preferably satisfies the relationship between the strand tensile modulus E and the crystallite size Lc (nm) given by formula (1), wherein the left-hand intercept of formula (1) is more preferably 135, and even more preferably 140. The right-hand intercept of formula (1) is more preferably 175, and even more preferably 170. 50 × Lc + 130 ≤ E ≤ 50 × Lc + 180 ... (1).

[0022] A carbon fiber bundle is a polycrystalline material composed of virtually countless graphite crystallites. Increasing the maximum temperature of the carbonization process increases the crystallinity of the crystallites. That is, rearrangement of the carbon network plane occurs, increasing the crystal size, and simultaneously, crystal orientation also progresses, so the strand tensile modulus E of the carbon fiber tends to increase. Therefore, a relationship can be observed between the strand tensile modulus E and the crystallite size Lc, as shown in equation (1). If the intercept of the left side of equation (1) is 130 or higher, the strand tensile modulus E can be efficiently improved even at a low carbonization temperature, which is preferable because it allows for a high strand tensile modulus E while suppressing fuzzing due to abrasion. If the intercept of the left side of equation (1) is 180 or lower, it is preferable because it eliminates the need to raise the maximum temperature of the carbonization process unnecessarily to increase the strand tensile modulus E, resulting in excellent abrasion resistance and suppression of entanglement due to ring-shaped fuzzing that occurs when the carbon fiber bundle is unwound for higher-order processing. The strand tensile modulus E and the crystallite size Lc can be measured by the method described above. To achieve this relationship between the strand tensile modulus E and crystallite size Lc, it can be controlled by appropriately controlling the application of tension to the fiber bundle during one of the heat treatment processes in the carbon fiber bundle manufacturing process, improving the double structure, or changing the carbonization temperature.

[0023] The carbon fiber bundle of the present invention has 40 or fewer fuzz particles per meter, preferably 35 or fewer, and more preferably 30 or fewer. Fuzz particles per meter refer to the fuzz present inside the carbon fiber bundle when the carbon fiber bundle is pulled out from the bobbin. If the number of fuzz particles per meter is 40 or fewer, the ring-shaped fuzz that occurs when the carbon fiber bundle is unwound for higher-level processing can be sufficiently suppressed. The method for measuring the number of fuzz particles in a carbon fiber bundle is to pull out 10m of the carbon fiber bundle from the bobbin, and then, using a force that does not generate fuzz, separate the carbon fiber bundle into individual fibers so that the thickness of the carbon fiber bundle is equivalent to two single fibers. If any fuzz is found, it is collected, its number is measured, and the number is calculated as the number per meter. At this time, fuzz generated during the stage of separating into individual fibers is excluded. To control the number of fuzz particles inherent in the carbon fiber bundle within a given range, this can be achieved by appropriately controlling the heat generation rate of the single fiber, the number of filaments N, the single fiber fineness, and the yarn width during the flame-retardant treatment process, as described later.

[0024] The carbon fiber bundle of the present invention has a carbon fiber bundle in which the proportion of fluff with a double-layered cross-section is 1 to 25%, preferably 1 to 20%, and more preferably 2 to 15%. Fuzz with a double-layered cross-section is defined as fluff that, when collected from the carbon fiber bundle and observed in cross-section with a scanning electron microscope (SEM), has a structure with an inner and outer layer as shown in Figure 1, or has a hole in the center of the fluff's cross-section as shown in Figure 2. However, fluff whose cross-section is judged to be fractured by bending as shown in Figure 3 is excluded from the total number of fluffs used to calculate the proportion of fluff with a double-layered cross-section among the fluffs contained in the carbon fiber bundle. This is because it is considered that the fluff was generated during the manufacturing process of the carbon fiber bundle and was not contained within the carbon fiber bundle, but rather fractured due to bending load applied when the fluff was collected.

[0025] While the mechanism by which the cross-sections of the fluff within these carbon fiber bundles have a double structure is not fully understood, it is thought to be as follows: When temperature variations occur during the flame-retardant process, the areas that become particularly hot develop a double structure that is larger than the average double structure within the flame-retardant fiber bundle. This leads to single fiber breakage of the carbon fiber bundle under weak loads, resulting in a cross-section with a double structure. From this, it is thought that fluff with a double structure is more likely to occur under weak loads and is considered to be the main cause of fluff within carbon fiber bundles.

[0026] When these particularly weak fibers are included in a certain proportion of the normally generated fibers, ring-shaped fibers are formed that wrap around other fibers, increasing the degree of wrapping. Therefore, it is thought that the presence of a certain proportion of fibers with a double-layered cross-section within the carbon fiber bundle, relative to the total amount of fibers, causes the wrapping of ring-shaped fibers to occur when the carbon fiber bundle is unwound for higher-level processing.

[0027] Therefore, if the proportion of fluff with a double-layered cross-section among the fluff contained in the carbon fiber bundle is 25% or less, the ring-shaped fluff that occurs when the carbon fiber bundle is unwound for higher-order processing can be sufficiently suppressed. If the proportion of fluff with a double-layered cross-section among the fluff contained in the carbon fiber bundle is 1% or more, the strand tensile strength of the carbon fiber bundle does not decrease, and the proportion of single fibers with low strength that affect the fluff is reduced, thereby suppressing the ring-shaped fluff that occurs near the roller when the carbon fiber bundle is unwound.

[0028] While commercially available carbon fiber bundles do not contain fluff with a double-layered cross-section, this can be controlled within a certain range by controlling the flame-retardant process, as described below. To determine whether a fluff within a carbon fiber bundle has a double-layered cross-section, the carbon fiber bundle wound on the bobbin is pulled out using the method described above, the fluff inside the bundle is collected, and the cross-section is observed using a scanning electron microscope (SEM) (details will be described later). Controlling the proportion of fluff within a carbon fiber bundle that has a double-layered cross-section within this range can be achieved by appropriately controlling the heat generation rate of the single fibers, the number of filaments N, the single fiber fineness, and the yarn width in the flame-retardant process, as described below.

[0029] In the carbon fiber bundle of the present invention, the proportion of cross-sections in which the area ratio of the fluff within the carbon fiber bundle that is perpendicular to the fiber axis is 50% or less is preferably 0 to 3%, more preferably 0.1 to 2.5%, and even more preferably 0.5 to 1.5%.

[0030] Here, the cross-section of the fluff within the carbon fiber bundle refers to the cross-section observed when the fluff inside the carbon fiber bundle is collected and the cross-section perpendicular to the fiber axis is observed using a scanning electron microscope (SEM). Furthermore, a cross-section in which the area ratio perpendicular to the fiber axis is 50% or less means that the cross-section of the fluff within the carbon fiber bundle is not substantially perpendicular, and has deformed as shown in Figure 4, or split from the center as shown in Figure 5, without retaining the shape of the original single fiber.

[0031] The proportion of the cross-sectional area of ​​fluff within a carbon fiber bundle is defined as the proportion of cross-sections where the area perpendicular to the fiber axis is 50% or less of the average cross-sectional area of ​​a cross-section obtained by cutting a single fiber perpendicularly. Therefore, as shown in Figure 5, even if the cross-section of the fluff is perpendicular, it may include sections that are split from the center. A cross-section perpendicular to the fiber axis is defined as a cross-section at an angle of 85 to 95 degrees with respect to the fiber axis.

[0032] The reason why the area ratio of the fluff cross-sections within the carbon fiber bundle that is perpendicular to the fiber axis is 50% or less is not clearly understood, but it is thought to be as follows: In other words, the temperature variation in the flame-retardant process is particularly large, and the double structure becomes even more specific than in the fluff cross-sections with a double structure, so that single fiber breakage of the carbon fiber bundle occurs even with a weaker load, and the cross-section of the fluff with a double structure is deformed. From this, it is thought that fluff with an area ratio of 50% or less of the area ratio of the fluff cross-sections within the carbon fiber bundle that is perpendicular to the fiber axis is more likely to occur with a weaker load than fluff with a double structure, and it is estimated that this leads to a significant increase in entanglement by ring-shaped fluff when the carbon fiber bundle is unwound for higher-order processing. For this reason, it is preferable that the proportion of the fluff cross-sections within the carbon fiber bundle that have an area ratio of 50% or less of the area ratio perpendicular to the fiber axis is 3% or less, as this can sufficiently suppress entanglement by ring-shaped fluff when the carbon fiber bundle is unwound for higher-order processing.

[0033] For fluff within a carbon fiber bundle, the area ratio of the cross-section perpendicular to the fiber axis is 50% or less. This is determined by pulling out the carbon fiber bundle wound on a bobbin as described above, collecting the fluff inside the carbon fiber bundle, and measuring the angle of the image obtained by observing the cross-section from an oblique angle of approximately 45° using an SEM. The area ratio of the cross-section that is 85-95° to the fiber axis is then extracted from the image analysis of the same cross-section observed from the front using an SEM. Furthermore, the cross-sectional area of ​​a single fiber in the carbon fiber bundle can be measured by cutting the carbon fiber bundle perpendicularly with a single blade to obtain a perpendicular cross-section, observing the cross-section of the extracted single fiber from the front using an SEM, and analyzing the image using image analysis software.

[0034] To control the area within the carbon fiber bundle where the proportion of the fluff cross-section perpendicular to the fiber axis is 50% or less, this can be achieved by appropriately controlling the heat generation rate of the single fiber, the number of filaments N, the single fiber fineness, and the yarn width in the flame-retardant process, as described later.

[0035] The yarn width W of the carbon fiber bundle of the present invention is preferably 5 to 8 mm, more preferably 6 to 8 mm, and even more preferably 7 to 8 mm. The yarn width W of the carbon fiber bundle is the width of the carbon fiber bundle when it is unwound from the bobbin, and unless a special fiber opening process is included, it generally reflects the width of the fiber bundle from the flame-retardant process. A yarn width W of 5 mm or more is preferable because it can suppress fuzzing caused by friction, thus suppressing fuzzing that occurs when the carbon fiber bundle is unwound. A yarn width W of 8 mm or less is preferable because it can suppress the generation of fuzzing caused by the carbon fiber bundle spreading more than necessary when it is unwound from the bobbin. The yarn width W of the carbon fiber bundle can be measured by unwinding the carbon fiber bundle from the bobbin and using a ruler or the like. Such a yarn width W of the carbon fiber bundle can be achieved by the yarn width of the polyacrylonitrile-based precursor fiber bundle and the tension of the flame-retardant fiber bundle in the flame-retardant process.

[0036] The number of filaments N in the carbon fiber bundle of the present invention is preferably 10,000 to 50,000, more preferably 10,000 to 30,000, and even more preferably 15,000 to 25,000. The number of filaments N in a carbon fiber bundle is the number of single fibers that make up the carbon fiber bundle. A number of filaments N of 10,000 or more is preferable because it can reduce the possibility of specific fluff inherent in the carbon fiber bundle coming out onto the surface of the carbon fiber bundle, and can sufficiently reduce fluff when the carbon fiber bundle is wound onto a bobbin. A number of filaments N of 50,000 or less is preferable because it can suppress fluff caused by friction, and can suppress fluff generated when the carbon fiber bundle is wound onto a bobbin. The number of filaments N in a carbon fiber bundle can be determined from the average single fiber diameter B of the carbon fiber bundle, the specific gravity of the carbon fiber bundle, and the basis weight (mass per unit length), which will be described later. The number of filaments N in such a carbon fiber bundle can be achieved by increasing the number of pores in the spinneret during the manufacturing process of the polyacrylonitrile precursor fiber bundle or by stacking multiple polyacrylonitrile precursor fiber bundles.

[0037] The knot strength A [MPa] of the carbon fiber bundle of the present invention is preferably -88B + 1360 ≤ A, more preferably -88B + 1370 ≤ A, and even more preferably -88B + 1390 ≤ A, in relation to the average single fiber diameter B (μm). Knot strength is an index that reflects the mechanical properties of the fiber bundle other than the fiber axis direction, and is a parameter that reflects the strength against bending loads and compressive loads applied from directions other than the fiber axis direction when the carbon fiber bundle is unwound from the bobbin. It is preferable that such knot strength satisfies -88B + 1360 ≤ A because it can reduce fluffing when the carbon fiber bundle is unwound from the bobbin. Such knot strength can be determined by the method for determining the knot strength of the carbon fiber bundle described later. To increase such knot strength of the carbon fiber bundle, it is preferable to appropriately control the heat generation rate of the single fiber, the number of filaments N, the single fiber fineness, and the yarn width in the flame-retardant process in the carbon fiber bundle manufacturing method of the present invention described later.

[0038] The carbon fiber bundle of the present invention preferably has a tensile strength (also abbreviated simply as strand tensile strength) of 5.5 to 7.0 GPa in a resin-impregnated strand tensile test, more preferably 5.8 to 6.8 GPa, and even more preferably 5.9 to 6.7 GPa. Strand tensile strength is a parameter that strongly relates to the average value of the single fiber strength and therefore also affects the tensile strength of the single fiber fluff; a higher value is preferable. However, the breadth of strength variation is more important than the average value of the single fiber strength. A strand tensile strength of 5.5 GPa or higher is preferable because it sufficiently suppresses entanglement due to ring-shaped fluff generated when the carbon fiber bundle is unwound for higher-order processing. While a higher strand tensile strength is preferable, a strand tensile strength of 7.0 GPa or lower is also preferable because it sufficiently suppresses entanglement due to ring-shaped fluff generated when the carbon fiber bundle is unwound for higher-order processing. The strand tensile strength can be determined by the method described later for the strand tensile test of the carbon fiber bundle. These parameters can be controlled by using the carbon fiber bundle manufacturing method of the present invention, which will be described later.

[0039] The carbon fiber bundle of the present invention preferably has an outer peripheral area ratio (outer layer area ratio) of the double structure to the entire cross-section perpendicular to the fiber axis of the fiber (hereinafter referred to as the outer layer area ratio), which is 85 to 95 area%, more preferably 87 to 94 area%, and even more preferably 89 to 93 area%. Here, the outer layer area ratio is the area ratio (%) obtained by dividing the area of ​​the outer peripheral area observed when observing the cross-section perpendicular to the fiber axis of a single fiber with an optical microscope by the entire cross-sectional area perpendicular to the fiber axis of the single fiber.

[0040] The interior of a single fiber has a lower degree of crystalline orientation and a lower strand tensile modulus E than the outer layer. Therefore, a higher ratio of this outer layer area is preferable because it suppresses fluffing (single fiber breakage). A ratio of 85% or more of this outer layer area is preferable because it can suppress specific fluffing inherent in the carbon fiber bundle. A ratio of 95% or less of this outer layer area is preferable because it can suppress fluffing caused by abrasion, which tends to occur due to excessive heat treatment in the flame-retardant process.

[0041] The outer layer area can be measured by embedding a carbon fiber bundle in resin, polishing a cross-section perpendicular to the fiber axis, and observing the cross-section with an optical microscope (details will be described later). This outer layer area ratio can be achieved by appropriately controlling the heat generation rate of the single fiber, the number of filaments N, the single fiber fineness, and the yarn width in the flame-retardant process, as described later.

[0042] The present invention addresses the problem of producing carbon fiber bundles that can suppress entanglement caused by ring-shaped fluff when the carbon fiber bundle is unwound for higher-order processing by suppressing specific fluff inherent in the carbon fiber bundle. By appropriately controlling the heat generation rate of the single fibers, the number of filaments N, the fineness of the single fibers, and the yarn width in the flame-retardant process, it is possible to ensure sufficient heat removal relative to the total heat generated in the flame-retardant fiber bundle, thereby reducing temperature variations within the flame-retardant fiber bundle. Preferred embodiments for carrying out this invention are described in detail below.

[0043] Preferably, a polyacrylonitrile polymer is used as the raw material for producing polyacrylonitrile precursor fiber bundles. In this invention, a polyacrylonitrile polymer refers to one in which at least acrylonitrile is the main component of the polymer backbone, and the main component usually refers to a component that accounts for 90 to 100% by mass of the polymer backbone.

[0044] In the production of polyacrylonitrile precursor fiber bundles, the polyacrylonitrile polymer preferably contains copolymer components such as itaconic acid, acrylamide, and methacrylic acid, from the viewpoint of improving spinnability and efficiently performing flame-retardant treatment. In the production of polyacrylonitrile precursor fiber bundles, the method for producing the polyacrylonitrile polymer can be selected from known polymerization methods.

[0045] In the production of a polyacrylonitrile-based precursor fiber bundle suitable for obtaining the carbon fiber bundle of the present invention, the spinning solution is obtained by dissolving the aforementioned polyacrylonitrile-based polymer in a polyacrylonitrile-soluble solvent such as dimethyl sulfoxide, dimethylformamide, dimethylacetamide, or an aqueous solution of nitrate, zinc chloride, and sodium rhodane.

[0046] There are no particular restrictions on the method for producing the polyacrylonitrile fiber bundle used in the present invention, but preferably wet spinning or wet-dry spinning is used, followed by processes such as stretching, washing, application of oil, drying and densification, and post-stretching if necessary. There are no particular restrictions on the number of holes in the spinneret in the production process of the polyacrylonitrile precursor fiber bundle, but in order to achieve the aforementioned number of filaments N of the carbon fiber bundle, the number of holes is preferably 1,000 to 10,000, considering the ease of splicing.

[0047] In the production of polyacrylonitrile precursor fiber bundles, it is preferable that the coagulation bath contains a solvent such as dimethyl sulfoxide, dimethylformamide, and dimethylacetamide, which were used as the solvent for the spinning solution, and a so-called coagulation promoting component. As the coagulation promoting component, a component that does not dissolve the aforementioned polyacrylonitrile polymer and is compatible with the solvent used in the spinning solution can be used. Preferably, water is used as the coagulation promoting component.

[0048] In the production of polyacrylonitrile precursor fiber bundles, it is preferable to use a multi-stage washing bath with a water bath temperature of 30 to 98°C during the washing process.

[0049] Furthermore, the stretching ratio in the water bath stretching process is preferably 2 to 6 times.

[0050] After the water bath stretching process, an oily agent, preferably made of silicone, is applied to the yarn to prevent adhesion between individual fibers. This silicone oily agent is preferably a modified silicone, and preferably contains a highly heat-resistant amino-modified silicone.

[0051] The drying heat treatment process can utilize known methods. For example, a drying temperature of 100-200°C is exemplified.

[0052] The dried yarn is preferably further stretched in pressurized steam or under dry heat, from the viewpoint of density and productivity of the resulting polyacrylonitrile-based precursor fiber bundles. The steam pressure or temperature and stretching ratio during the stretching process should be appropriately selected within a range that does not cause yarn breakage or fluffing.

[0053] In the carbon fiber bundle manufacturing method of the present invention, the single fiber fineness of the polyacrylonitrile-based precursor fiber bundle is 0.9 to 2.2 dtex, preferably 1.0 to 1.8 dtex, and more preferably 1.1 to 1.7 dtex. The single fiber fineness of the polyacrylonitrile-based precursor fiber bundle refers to the diameter of the single fiber in the polyacrylonitrile-based precursor fiber bundle. If the single fiber fineness of the polyacrylonitrile-based precursor fiber bundle is 0.9 dtex or higher, the abrasion resistance of the resulting carbon fiber bundle is improved, and the fuzzing that occurs when the carbon fiber bundle is unwound can be suppressed. If the single fiber fineness of the polyacrylonitrile-based precursor fiber bundle is 2.2 dtex or lower, sufficient heat removal can be secured relative to the total heat generated in the flame-retardant fiber bundle during the flame-retardant process, temperature variations within the flame-retardant fiber bundle can be reduced, and specific fuzzing inherent in the carbon fiber bundle can be suppressed. The single fiber fineness of a polyacrylonitrile precursor fiber bundle can be calculated from the mass, density, and number of filaments (N) per unit length of the polyacrylonitrile precursor fiber bundle. Such a polyacrylonitrile precursor fiber bundle can be achieved by controlling the extrusion rate and the stretching ratio at each stage of the manufacturing process.

[0054] In the production of carbon fiber bundles of the present invention, following the polyacrylonitrile-based precursor fiber bundle production step, preferably a filamentation process is performed before the flame-retardant step, depending on the number of filaments N of the polyacrylonitrile-based precursor fiber bundle. A preferred form of filamentation is to unwind the polyacrylonitrile-based precursor fiber bundle from the creel and then perform filamentation according to the number of filaments N of the polyacrylonitrile-based precursor fiber bundle so that the number of filaments N of the target carbon fiber bundle is determined.

[0055] In the carbon fiber bundle manufacturing method of the present invention, the temperature in the step of heat-treating the polyacrylonitrile-based precursor fiber bundle in an oxidizing atmosphere (flame-retardant step) is 200 to 300°C, preferably 220 to 290°C, and more preferably 230 to 280°C. If the heat treatment temperature is 200°C or higher, it is less likely that untreated areas will form within the flame-retardant fiber bundle due to a low heat treatment temperature, thus reducing the occurrence of uneven double-layer structures, and thus significantly reducing fluffing when the carbon fiber bundle is wound onto a bobbin. If the heat treatment temperature is 300°C or lower, the heat generation rate will not be unnecessarily high, thus reducing temperature variations within the flame-retardant fiber bundle and suppressing specific fluffing inherent in the carbon fiber bundle. To measure the heat treatment temperature, a thermometer such as a thermocouple can be inserted into the heat treatment furnace of the flame-retardant step to measure the furnace temperature. If temperature variations or temperature distributions are observed when measuring the furnace temperature at several points, the simple average temperature can be calculated.

[0056] The present invention provides a method for producing carbon fiber bundles in which, when the heat generation rate of a single fiber in the flame-retardant process is q (J / g / s), the number of filaments N is N (threads), the fineness of the single fiber of the flame-retardant fiber bundle is d (dtex), and the yarn width is W (mm), the density is 1.22 to 1.24 g / cm³. 3 The heat generation rate Q, calculated using equation (3), is 150-500 J / m³ until it reaches this point. 2 The heat treatment is performed so that the temperature becomes / s. Q = q × N × d / W / 10 ... (3).

[0057] The density of flame-retardant fiber bundles is commonly used as an indicator of the progress of the flame-retardant reaction. Density: 1.22–1.24 g / cm³ 3 This means that it is the initial stage of the flame-retardant process, and controlling the heat generation rate in this initial stage of the flame-retardant process within an appropriate range is important because it allows control of the proportion of fluff in the carbon fiber bundle that has a double-layered cross-section, and suppresses entanglement caused by ring-shaped fluff that occurs when the carbon fiber bundle is unwound for higher-level processing.

[0058] The density is 1.22 g / cm³. 3If it is as described above, in the subsequent flame-resistant treatment process, even when heat treatment is performed at a high temperature, the rapid increase in the heat generation rate within the flame-resistant fiber bundle can be suppressed, leading to the suppression of temperature spots within the flame-resistant fiber bundle, and the proportion of the hairs within the carbon fiber bundle having a double structure in cross-section can be suppressed.

[0059] If such density is 1.24 g / cm 3 If it is below, it is a structure capable of sufficiently controlling the double structure of the flame-resistant fiber bundle, and when controlling the heat generation rate described later, the effect of suppressing the proportion of the hairs within the carbon fiber bundle having a double structure in cross-section can be sufficiently enhanced.

[0060] To confirm that it is within the range where such density is subjected to heat treatment at the heat generation rate Q described later, it is sufficient to collect the fiber bundle during the flame-resistant treatment process and measure the density (the density measurement method will be described later). For example, if the density of the flame-resistant fiber bundle is lower than the specified value, the density can be adjusted by increasing the temperature or lengthening the flame-resistant treatment time. Here, the oxidizing atmosphere refers to an atmosphere containing 10% by mass or more of known oxidizing substances such as oxygen and nitrogen dioxide, and the air atmosphere is preferred for simplicity.

[0061] In the flame-resistant treatment process of the method for manufacturing a carbon fiber bundle of the present invention, the heat generation rate Q until the density reaches 1.22 to 1.24 g / cm 3 is 150 to 500 J / m 2 / s, preferably 160 to 400 J / m 2 / s, more preferably 180 to 350 J / m 2 / s. Note that as long as it is after being controlled within the range where the heat generation rate Q until the density reaches 1.22 to 1.24 g / cm 3 is applied, even if the heat generation rate Q until the next set density range is changed, it is considered that this requirement is satisfied. For example, if the heat generation rate Q until the density reaches 1.23 g / cm 3 is controlled within the range of 150 to 500 J / m 2 / s, the heat generation rate Q until the density becomes larger than 1.23 g / cm 3 may be outside the range of 150 to 500 J / m 2 / s.

[0062] In this invention, the heat generation rate Q is calculated by dividing the total heat generation rate per unit length of the flame-resistant fiber bundle (numerator of equation (3)) by the yarn width of the flame-resistant fiber bundle. This represents the heat generation rate per unit area of ​​the flame-resistant fiber bundle and signifies the relationship between heat generation and heat removal in the flame-resistant fiber bundle; in other words, it is a heat generation rate that takes into account the effect of heat removal. Therefore, since the heat generation rate Q is a parameter that reflects the temperature variation of the flame-resistant fiber bundle, a small heat generation rate Q means that the temperature variation of the flame-resistant fiber bundle is small.

[0063] The resulting heat generation rate Q is 150 J / m². 2 If the temperature is above / s, the balance between heat generation and heat removal is good, making it less likely for untreated areas to form within the flame-resistant fiber bundle, eliminating unevenness in the double-layer structure, and thus significantly reducing fluffing when the carbon fiber bundle is wound onto a bobbin.

[0064] The resulting heat generation rate Q is 500 J / m 2 If the temperature is less than / s, the heat removal rate is sufficiently large compared to the heat generation rate, which can reduce temperature variations within the flame-resistant fiber bundle and suppress specific fluffing inherent in the carbon fiber bundle.

[0065] To calculate the heat generation rate Q, the heat generation rate q (J / g / s) of a single fiber and the yarn width W (mm) are measured using the method described later. The rate can then be calculated from equation (3) using the number of filaments N (threads) and the single fiber fineness d (dtex) of the flame-resistant fiber bundle. This heat generation rate Q can be controlled by the heat treatment temperature in the flame-resistant process, the number of filaments N, the single fiber fineness of the flame-resistant fiber bundle, and the pitch (width) of the roller grooves.

[0066] In the method for producing carbon fiber bundles according to the present invention, the final density of the flame-resistant fiber bundle in the flame-retardant step is 1.38 to 1.50 g / cm³. 3 The density is preferably 1.42 to 1.48 g / cm³. 3 The final density of the flame-resistant fiber bundle is 1.38 g / cm³. 3If the above conditions are met, fuzzing caused by friction of the carbon fiber bundle can be suppressed, thus suppressing the fuzzing that occurs when the carbon fiber bundle is unwound. The final density of the flame-resistant fiber bundle is 1.50 g / cm³. 3 If the following conditions are met, excessive heat treatment can be prevented, thereby reducing the proportion of fluff within the carbon fiber bundle that has a double-layered cross-section, and thus suppressing the fluff generated when the carbon fiber bundle is unwound.

[0067] To confirm that the final density of the flame-resistant fiber bundle is within the specified range, one can take a sample of the flame-resistant fiber bundle and measure its density (the method for measuring density will be described later). For example, if the density of the flame-resistant fiber bundle is lower than specified, the density can be adjusted by increasing the temperature or extending the flame-retardant treatment time. Here, an oxidizing atmosphere is an atmosphere containing 10% by mass or more of a known oxidizing substance such as oxygen or nitrogen dioxide, and an air atmosphere is preferred for its simplicity.

[0068] In the flame-retardant step of the carbon fiber bundle manufacturing method of the present invention, when the heat generation rate of a single fiber is q (J / g / s), the number of filaments N is N (threads), the fineness of the single fiber of the flame-retardant fiber bundle is d (dtex), and the yarn width is W (mm), the density is 1.22 to 1.24 g / cm³. 3 The heat generation rate Q, calculated using equation (3), is 150-500 J / m³ until it reaches this point. 2 After heat treatment to achieve a density of 1.32-1.35 g / cm³, 3 The heat generation rate Q, calculated using equation (3), is preferably 300 to 1200 J / m² until the value is reached. 2 The heat treatment is performed to achieve a temperature of / s, more preferably 400-1100 J / m 2 The heat treatment is performed to achieve a temperature of / s, and more preferably 500-1000 J / m 2 Heat treatment is performed to achieve a temperature of / s.

[0069] For example, a density of 1.23 g / cm³ 3 The heat generation rate Q is 150-500 J / m² until it reaches this point. 2 After heat treatment to achieve a density of / s, the density becomes 1.33 g / cm³. 3 The heat generation rate Q is 300-1200 J / m² until it reaches this point. 2This requirement is met when the heat treatment is performed to achieve a value of / s.

[0070] However, the density at this time is 1.23 g / cm³. 3 The heat generation rate Q is 150-500 J / m² until it reaches this point. 2 After heat treatment to achieve a density of 1.24 g / cm³, 3 The heat generation rate Q is 150-500 J / m² until it reaches this point. 2 Although not in / s, the heat generation rate Q is 300-1200 J / m 2 Make it so that it becomes / s (for example, 800 J / m 2 Heat treatment is preferable, but any heat generation rate Q outside the range of (e.g., 1500 J / m²) is not recommended. 2 It is not preferable to use a density of 1.32-1.35 g / cm³. 3 The flame-retardant fiber bundles exhibit a moderate degree of flame-retardant reaction. The heat generation rate at a moderate degree of flame-retardant reaction can affect the final double structure of the flame-retardant fiber bundles and carbon fiber bundles. Therefore, the density should be 1.22-1.24 g / cm³. 3 Density 1.32~1.35 g / cm³ 3 It is preferable to control the heat generation rate Q until it reaches this point within a certain range.

[0071] If the heat generation rate Q is 300 or higher, the balance between the amount of heat generated and the amount of heat removed is good, making it less likely for untreated areas to form within the flame-resistant fiber bundle, and eliminating unevenness in the double-layer structure. This is preferable because it can sufficiently reduce fluff when the carbon fiber bundle is wound onto a bobbin.

[0072] If the heat generation rate Q is 1200 or less, the heat removal rate is sufficiently large compared to the heat generation rate, which reduces temperature variations within the flame-resistant fiber bundle and suppresses specific fluffiness inherent in the carbon fiber bundle, making it preferable.

[0073] To confirm that the density is within the range for heat treatment at a heat generation rate Q, one can take a fiber bundle during the flame-retardant treatment process and measure its density (the method for measuring density will be described later). For example, if the density of the flame-retardant fiber bundle is lower than specified, the density can be adjusted by increasing the temperature or extending the flame-retardant treatment time.

[0074] Here, an oxidizing atmosphere refers to an atmosphere containing 10% by mass or more of a known oxidizing substance such as oxygen or nitrogen dioxide, and an air atmosphere is preferred for simplicity. To calculate the heat generation rate Q, the heat generation rate q (J / g / s) of the single fiber, the number of filaments N (threads), the single fiber fineness d (dtex) of the flame-resistant fiber bundle, and the yarn width W (mm) are measured using the method described later, and the rate can be calculated from equation (3). This heat generation rate Q can be controlled by the heat treatment temperature in the flame-resistant process, the number of filaments N, the single fiber fineness of the flame-resistant fiber bundle, and the pitch (width) of the roller grooves.

[0075] The flame-retardant step in the carbon fiber bundle manufacturing method of the present invention involves a density of 1.22 to 1.24 g / cm³. 3 The heat generation rate Q, calculated using equation (3), is 150 to 500 J / m³ until it reaches this value. 2 After heat treatment to achieve a density of 1.32-1.35 g / cm³, 3 The heat generation rate Q, calculated using equation (3), is 300 to 1200 J / m² until it reaches this value. 2 The material is heat-treated to achieve a density of 1.38-1.50 g / cm³, and then heat-treated to achieve a density of 1.38-1.50 g / cm³. 3 The heat generation rate Q, calculated using equation (3), is preferably 900 to 1500 J / m² until the value is reached. 2 / s, more preferably 1000~1400 J / m 2 / s, more preferably 1100-1300 J / m 2 Heat treatment is performed to achieve a temperature of / s.

[0076] For example, a density of 1.23 g / cm³ 3 The heat generation rate Q is 150-500 J / m 2 After heat treatment at / s, the density is 1.33 g / cm³. 3 The heat generation rate Q is 300-1200 J / m 2 The heat treatment was performed at / s, and then further 1.48 g / cm³.3 The heat generation rate Q, which can be calculated using equation (3), is 900 to 1500 J / m³. 2 This requirement is met when heat treatment is performed as / s.

[0077] However, the density at this time is 1.33 g / cm³. 3 The heat generation rate Q is 150-500 J / m 2 After heat treatment at / s, the density becomes 1.35 g / cm³. 3 The conditions for heat treatment up to the aforementioned preferred heat release rate Q(300~1200 J / m²) 2 The rate of heat generation Q is not (not / s), but rather (900~1500 J / m). 2 It is within the range of / s, for example, 1250 J / m 2 Heat treatment at a rate of Q (e.g., 1600 J / m²) is preferable, but any rate of heat generation outside the range of Q (e.g., 1600 J / m²) is not preferable. 2 Using / s is not advisable.

[0078] Furthermore, the density is 1.38-1.50 g / cm³. 3 Since this is the final density of the flame-retardant fiber bundle in the present invention, it may affect the double structure of the carbon fiber bundle, therefore the density should be 1.38~1.50 g / cm³. 3 Within this range, the heat generation rate Q is 900-1500 J / m 2 It is preferable to keep the density within the range of / s. For example, a density of 1.38 g / cm³. 3 The heat generation rate Q is 900-1500 J / m 2 / s (for example, 1000 J / m 2 After setting it to / s, it is further set to 1.50 g / cm³ 3 Even when heat treatment is performed to this extent, the heat generation rate Q should be 900-1500 J / m 2 / s (for example, 1400 J / m 2 It is preferable to use / s).

[0079] Density 1.32~1.35g / cm 3 The flame-retardant fiber bundle showed a moderate degree of flame-retardant reaction, and subsequently its density was 1.38-1.50 g / cm³. 3The heat generation rate during the heat treatment process until the desired density is reached can affect the double structure of the final flame-resistant fiber bundles and carbon fiber bundles, therefore the density should be 1.32-1.35 g / cm³. 3 Density 1.38~1.50 g / cm³ 3 It is preferable to control the heat generation rate Q until it reaches this point within a certain range.

[0080] The resulting heat generation rate Q is 900 J / m 2 A value of / s or higher is preferable because it provides a good balance between heat generation and heat removal, making it less likely for untreated areas to form within the flame-resistant fiber bundle, eliminating unevenness in the double-layer structure, and thus significantly reducing fluffing when the carbon fiber bundle is wound onto a bobbin.

[0081] The resulting heat generation rate Q is 1500 J / m². 2 If the temperature is less than or equal to / s, the heat removal rate is sufficiently large compared to the heat generation rate, which reduces temperature variations within the flame-resistant fiber bundle and suppresses certain fluffiness inherent in the carbon fiber bundle, making it preferable.

[0082] To confirm that the density is within the range for heat treatment at a heat generation rate Q, one can take a fiber bundle during the flame-retardant treatment process and measure its density (the method for measuring density will be described later). For example, if the density of the flame-retardant fiber bundle is lower than specified, the density can be adjusted by increasing the temperature or extending the flame-retardant treatment time.

[0083] Here, an oxidizing atmosphere refers to an atmosphere containing 10% by mass or more of a known oxidizing substance such as oxygen or nitrogen dioxide, and an air atmosphere is preferred for simplicity. To calculate the heat generation rate Q, the heat generation rate q (J / g / s) of the single fiber, the number of filaments N (threads), the single fiber fineness d (dtex) of the flame-resistant fiber bundle, and the yarn width W (mm) are measured using the method described later, and the rate can be calculated from equation (3). This heat generation rate Q can be controlled by the heat treatment temperature in the flame-resistant process, the number of filaments N, the single fiber fineness of the flame-resistant fiber bundle, and the pitch (width) of the roller grooves.

[0084] In this invention, the density is 1.22 to 1.24 g / cm³. 3The heat generation rate Q, calculated using equation (3), is 150 to 500 J / m³ until it reaches this value. 2 After heat treatment to achieve a density of 1.38-1.50 g / cm³, the density is 1.38-1.50 g / cm³. 3 The tension applied to the flame-resistant fiber bundle during heat treatment until it reaches the desired state is 1.6 to 4.0 mN / dtex, preferably 2.5 to 4.0 mN / dtex, and more preferably 3.0 to 4.0 mN / dtex. For example, density 1.23 g / cm³ 3 After heat treatment, the density is 1.40 g / cm³. 3 The heat treatment must be performed with a tension that reaches the desired state. If the tension is 1.6 mN / dtex or higher, the orientation of the carbon fiber bundle is sufficiently increased, improving the strand tensile modulus E, and thus significantly reducing fluffing when the carbon fiber bundle is unwound from the bobbin. If the tension is 4.0 mN / dtex or lower, the fluffing inherent in the carbon fiber bundle can be suppressed. The tension applied to the flame-retardant fiber bundle in the flame-retardant process shall be expressed as the value obtained by dividing the tension (mN) measured at the exit of the flame-retardant furnace by the fineness (dtex) of the polyacrylonitrile-based precursor fiber bundle when completely dry.

[0085] In the production of carbon fiber bundles according to the present invention, the rollers in the flame-retardant process preferably have grooves in order to achieve the desired yarn width W of the carbon fiber bundle. The pitch (width) of the grooves on the rollers can be set according to the desired yarn width, and is preferably 5 to 8 mm.

[0086] In the production of carbon fiber bundles of the present invention, it is preferable to perform pre-carbonization following the polyacrylonitrile-based precursor fiber bundle production step and the flame retardation step. In the pre-carbonization step, the obtained flame retarded fiber bundle is subjected to a process in an inert atmosphere at a maximum temperature of 500 to 1,200°C, preferably with a density of 1.5 to 1.8 g / cm³. 3 Heat treatment until it reaches this state.

[0087] Following the aforementioned pre-carbonization, carbonization is carried out. In the present invention, in the carbonization process, the obtained pre-carbonized fiber bundles are manufactured in an inert atmosphere at a maximum temperature of 1,200 to 1,600°C. If the maximum temperature is 1,200°C or higher, it is possible to suppress entanglement due to ring-shaped fluff that occurs when the carbon fiber bundles are unwound for higher-level processing. If the maximum temperature is 1,600°C or lower, it is possible to suppress fluffing caused by friction of the carbon fiber bundles, thereby suppressing fluffing that occurs when the carbon fiber bundles are unwound.

[0088] The carbon fiber bundles obtained as described above are subjected to oxidation treatment to improve adhesion to the matrix resin, thereby introducing oxygen-containing functional groups. While gas-phase oxidation, liquid-phase oxidation, and liquid-phase electrolytic oxidation can be used as oxidation treatment methods, liquid-phase electrolytic oxidation is preferred from the viewpoint of high productivity and uniform treatment. There are no specific requirements for the liquid-phase electrolytic oxidation method; any known method may be used.

[0089] After such electrolytic treatment, a sizing treatment can be performed to impart bundle-forming properties to the resulting carbon fiber bundle. Depending on the type of matrix resin used in the composite material, a sizing agent with good compatibility with the matrix resin can be appropriately selected.

[0090] The methods for measuring the various physical properties described herein are as follows:

[0091] <Measurement of crystallite size Lc> The carbon fibers to be measured are aligned, and the measurement is performed using a wide-angle X-ray diffractometer under the following conditions. • X-ray source: CuKα rays (tube voltage 40kV, tube current 30mA) • Detectors: Goniometer + Monochromator + Scintillation Counter • Scanning range: 2θ = 10~40° • Scanning mode: Step scan, step unit 0.01°, scan speed 1° / min.

[0092] In the obtained diffraction pattern, for the peak appearing around 2θ = 25 to 26°, the full width at half maximum is determined, and from this value, the crystallite size is calculated using the following Scherrer's formula.

[0093] Crystallite size (nm) = Kλ / β0cosθ B However, K: 1.0, λ: 0.15418 nm (wavelength of X-ray) β0: (β E 2 -β1 2 ) 1 / 2 β E : Apparent full width at half maximum (measured value) rad, β1: 1.046×10 -2 rad θ B : Bragg diffraction angle.

[0094] <Tensile strength and tensile modulus E of strands in carbon fiber bundles> The tensile strength and tensile modulus E of strands in carbon fiber bundles are determined according to the resin-impregnated strand test method of JIS-R-7608 (2004) following the steps below. As the 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) is used, and as the curing conditions, normal pressure, a temperature of 125°C, and a time of 30 minutes are used. Ten resin-impregnated strands of the carbon fiber bundle are measured, and the average value is taken as the tensile strength of the strand. Strain is evaluated using an extensometer. The strain range is 0.1 to 0.6%.

[0095] <Number of protruding fibers inherent in carbon fiber bundles> The carbon fiber bundle is pulled out from the bobbin without tension for 10 m, and the carbon fiber bundle is divided into individual fibers so that the thickness of the carbon fiber bundle becomes two single fibers with a force that does not generate protruding fibers. If there are protruding fibers, they are collected and the number is measured. The number per meter is calculated as the number of protruding fibers inherent in the carbon fiber bundle from the measured number. At this time, the protruding fibers generated at the stage of dividing into individual fibers are excluded from the number.

[0096] <Percentage of fluff within a carbon fiber bundle that has a double-layered cross-section> Carbon fiber bundles are unwinded from bobbins without tension, and 50 fluffs are randomly collected from within the carbon fiber bundles. The tips of the collected fluffs are observed from the front and at approximately a 45° angle using a scanning electron microscope (SEM) "S-4800" manufactured by Hitachi High-Technologies Corporation. Among the observed cross-sections, those with a structure that appears as two concentric layers as shown in Figure 1, or those with a hole in the center of the fluff's cross-section as shown in Figure 2, are defined as having a "double structure." Fuzzes whose cross-sections are judged to have been fractured by bending, as shown in Figure 3, are excluded from the total number of fluffs because they are thought to have been generated during the carbon fiber bundle manufacturing process and not to have been inherent in the carbon fiber bundle, but rather fractured by bending load during fluff collection. If the total number of excluded fluffs is 5 or more, an additional 15 fluffs are collected to bring the total number of fluffs to 100 or more. The ratio of the total number of "cross-sections with a double structure" to the total number of fluffs other than those broken by bending, obtained in this way, is defined as the proportion of fluffs with a double structure in cross-section among the fluffs inherent in the carbon fiber bundle.

[0097] <Percentage of fluff within a carbon fiber bundle where the area ratio of the cross-section perpendicular to the fiber axis is 50% or less> The cross-sectional area of ​​a single fiber in a carbon fiber bundle is determined by cutting the single fiber with a single blade to obtain 30 perpendicular cross-sections, and then imaging the obtained cross-sections from the front using a Hitachi High-Technologies Corporation scanning electron microscope (SEM) "S-4800". The major axis of the obtained SEM images is measured using the ruler tool of the free image analysis software "Image J", and the average value of the 30 measurements is taken as the average cross-sectional area of ​​a single fiber in the carbon fiber bundle. Furthermore, for the SEM image of the "cross-section of the fluff inherent in the carbon fiber bundle" described above, an area at an angle of 85 to 95 degrees with respect to the fiber axis is selected using the protractor tool of the free image analysis software "Image J" for the image observed from approximately 45° obliquely. The area of ​​the cross-section perpendicular to the fiber axis is calculated by calculating the area of ​​the same cross-section imaged from the front using the free image analysis software "Image J" for the selected area. The ratio of the area of ​​the cross-section perpendicular to the fiber axis to the average cross-sectional area of ​​the single fibers of the carbon fiber bundle obtained above is calculated, and if it is 50% or less, it is defined as "fibers in the carbon fiber bundle where the area ratio of the cross-section perpendicular to the fiber axis is 50% or less." The ratio of the total number of "fibers in the carbon fiber bundle where the area ratio of the cross-section perpendicular to the fiber axis is 50% or less" to the total number of fibers other than those broken by bending, as determined by the method described above, is defined as the proportion of fibers in the carbon fiber bundle where the area ratio of the cross-section perpendicular to the fiber axis is 50% or less.

[0098] <Width W of carbon fiber bundle> The carbon fiber bundle is unwound from the bobbin without tension, ensuring that the bundle does not slacken, and the thread width is measured with a ruler. Three measurements are taken at 1m intervals, and the average value is used as the thread width W of the carbon fiber bundle.

[0099] <Knotting strength of carbon fiber bundles> A 150 mm long carbon fiber bundle is prepared as a test specimen by attaching 25 mm long gripping sections to both ends. When preparing the test specimen, 0.1 × 10 -3Apply a load of N / denier to align the carbon fiber bundles. Make one knot at the midpoint of the test piece, and conduct a bundle tensile test with a crosshead speed of 100 mm / min during tension. The measurement is performed on a total of 12 fiber bundles. Use the average value of 10 bundles excluding the maximum and minimum values as the measured value, and use the standard deviation of the 10 bundles as the standard deviation of the knot strength. For the knot strength, use the value obtained by dividing the maximum load value obtained in the bundle tensile test by the average cross-sectional area value of the carbon fiber bundle.

[0100] <Density measurement> Collect 1.0 - 3.0 g of flame-resistant fiber bundles and dry them to a constant weight at 120°C for 2 hours. Next, after measuring the dry weight C (g), impregnate them with ethanol and degas them sufficiently, then measure the fiber weight D (g) in the ethanol solvent bath, and calculate the fiber specific gravity using fiber specific gravity = (C × ρ) / (C - D). ρ is the specific gravity of ethanol at the measurement temperature.

[0101] <Average single fiber diameter B of carbon fiber bundle> For a carbon fiber bundle composed of a large number of carbon filaments to be measured, the mass A per unit length f (g / m) and density ρ (g / cm 3 ) are determined. Let the number of filaments N of the carbon fiber bundle to be measured be C f , and calculate the average single fiber diameter B (μm) of the carbon fiber bundle using the following formula. Average single fiber diameter B (μm) of carbon fiber bundle = ((A f / ρ / C f ) / π) (1 / 2) × 2 × 10 3 .

[0102] <Ratio of outer layer area to the entire cross-section perpendicular to the fiber axis of carbon fiber single fiber> The carbon fiber bundle to be measured is embedded in resin, the cross-section perpendicular to the fiber axis is polished, and the cross-section is observed at a total magnification of 1,000x using a 100x objective lens of an optical microscope. The outer layer area of ​​the double structure is measured from the cross-sectional microscope image of the polished surface. The analysis is performed using the image analysis software ImageJ. First, the cross-sectional image of a single fiber is binarized to separate the black and white regions. For the brightness distribution within the cross-section of the single fiber, the average value of the distribution is set as the threshold, and binarization is performed. The obtained binarized image is measured as the shortest distance from a point on the surface to the region with a line from black to white, with respect to the direction of the fiber diameter. This is measured for 5 points within the circumference of the same single fiber, and the average value is calculated as the outer layer thickness at that level. From the above, the area ratio (%) of the outer layer to the entire cross-section perpendicular to the fiber axis of the carbon fiber single fiber is calculated, and the average of 50 cross-sections is taken as the outer layer area ratio to the entire cross-section perpendicular to the fiber axis of the carbon fiber single fiber.

[0103] <heat generation rate q of a single fiber> Polyacrylonitrile precursor fiber bundles are dried at 120°C for 1 hour under reduced pressure of 10 mmHg or less, and then subjected to calorific value analysis. 2 mg of the dried polyacrylonitrile precursor fiber bundle is weighed into an aluminum sample pan. The aluminum sample pan is left uncovered, and measurements are taken from room temperature to 300°C using a thermal flux differential scanning calorimeter (Bruker AXS DSC3100SA) under a heating rate of 10°C / min and an air supply of 100 mL / min. The obtained data are used with the exothermic rate at 150°C set to zero, and the exothermic rate at a given temperature set to q.

[0104] <Quality of carbon fiber bundles during winding> A bobbin of carbon fiber bundles is placed on a creel, and under a tension of 1.6 mN / dtex, it is pulled up by a roller at 10 m / min and wound up by a winder. During this time, the amount of fuzz generated between the creel and the roller is counted for 10 minutes and evaluated using the following indicators.

[0105] A: 1~2 pieces / 10 minutes B: 3~5 pieces / 10 minutes C: 6 pieces~ / 10 minutes [Examples]

[0106] The present invention will be described in more detail below with reference to examples. However, the present invention is not limited to these examples. The measurement methods in these examples are as described above.

[0107] (Example 1) A copolymer consisting of acrylonitrile and itaconic acid was polymerized by solution polymerization using dimethyl sulfoxide as a solvent to produce a polyacrylonitrile copolymer and obtain a spinning stock. The obtained spinning stock was discharged into the air from a spinning die and then introduced into a coagulation bath consisting of a 35% aqueous solution of dimethyl sulfoxide controlled at 3°C ​​to obtain fiber bundles coagulated by a wet-dry spinning method. These fiber bundles were washed and stretched by conventional methods at 30-98°C. Subsequently, an amino-modified silicone-based silicone oil was applied to the fiber bundles after stretching in the water bath, and a drying and densification treatment was performed using a heated roller at 160°C to obtain polyacrylonitrile precursor fiber bundles with 12,000 single fibers. These bundles were then stretched 3.7 times in pressurized steam to obtain a total spinning stretch ratio of 13 times, resulting in polyacrylonitrile precursor fiber bundles with 12,000 single fibers. The amount of spinning solution discharged from the spinneret was adjusted so that the single fiber fineness of the polyacrylonitrile precursor fiber bundles was as shown in Table 2. The heat generation rate q of the single fibers of the obtained polyacrylonitrile precursor fiber bundles was measured using the method described above. Next, using the heat treatment temperature and flame-retardant treatment time conditions shown in Table 2, the polyacrylonitrile precursor fiber bundles were heat-treated in an air atmosphere oven at a draw ratio of 1 to obtain flame-retardant fiber bundles.

[0108] The obtained flame-resistant fiber bundles were subjected to a preliminary carbonization treatment in a nitrogen atmosphere at a temperature of 300-800°C to obtain preliminary carbonized fiber bundles. The obtained preliminary carbonized fiber bundles were then subjected to a carbonization treatment in a nitrogen atmosphere at a maximum temperature of 1,350°C. The obtained carbon fiber bundles were then subjected to surface treatment and sizing agent application to obtain the final carbon fiber bundles.

[0109] Table 1 shows the average single fiber diameter B, strand tensile modulus E, crystallite size Lc, strand tensile strength, yarn width W, number of filaments N, outer layer area ratio, knot strength A, number of filaments in the carbon fiber bundle, cross-section of the filaments in the carbon fiber bundle, and quality of the carbon fiber bundle at the time of winding. The number of filaments in the carbon fiber bundle was 38 / m, and 24% of the filaments had a double-layered cross-section, indicating good quality of the carbon fiber bundle at the time of winding.

[0110] (Example 2) Density 1.22~1.24g / cm 3 Except for setting the heat treatment temperature to 235°C, the procedure was the same as in Example 1. The number of fluff particles in the carbon fiber bundle was 3 particles / m, and 4% of the fluff particles had a double-layered cross-section. The quality of the carbon fiber bundle when unwound was very good. The obtained evaluation results are shown in Tables 1 and 2.

[0111] (Example 3) Thread width W: 5mm, Filament count N: 24,000, Density: 1.22~1.24g / cm² 3 The heat treatment temperature until this is achieved is 230°C, and the density is 1.38-1.50 g / cm³. 3 Except for setting the heat treatment temperature to 265°C, the procedure was the same as in Example 1. The number of fluff particles in the carbon fiber bundle was 32 particles / m, and 12% of the fluff particles had a double-layered cross-section. The quality of the carbon fiber bundle when unwound was very good. The obtained evaluation results are shown in Tables 1 and 2.

[0112] (Example 4) Yarn width W is 8mm, density 1.22~1.24g / cm 3 Except for setting the heat treatment temperature to 235°C, the procedure was the same as in Example 3. The number of fluff particles in the carbon fiber bundle was 2 particles / m, and 2% of the fluff particles had a double-layered cross-section. The quality of the carbon fiber bundle when unwound was very good. The obtained evaluation results are shown in Tables 1 and 2.

[0113] (Example 5) Density 1.22~1.24g / cm 3 Except for setting the heat treatment temperature to 235°C, the procedure was the same as in Example 1. The number of fluff particles in the carbon fiber bundle was 4 particles / m, and 2% of the fluff particles had a double-layered cross-section. The quality of the carbon fiber bundle when unwound was very good. The obtained evaluation results are shown in Tables 1 and 2.

[0114] (Example 6) Except for setting the single fiber fineness d of the flame-resistant fiber bundle to 1.2 dtex, the procedure was the same as in Example 5. The number of fluff particles in the carbon fiber bundle was 3 particles / m, and 4% of the fluff particles had a double-layered cross-section, indicating good quality when the carbon fiber bundle was unwound. The obtained evaluation results are shown in Tables 1 and 2.

[0115] (Example 7) Except for setting the yarn width W to 7 mm and the single fiber fineness d of the flame-resistant fiber bundle to 0.9 dtex, the procedure was the same as in Example 5. The number of fluff particles in the carbon fiber bundle was 5 particles / m, and 1% of the fluff particles had a double-layered cross-section, indicating good quality when the carbon fiber bundle was unwound. The obtained evaluation results are shown in Tables 1 and 2.

[0116] (Example 8) Density 1.38~1.50g / cm 3 Except for the tension of the flame-resistant fiber bundle during heat treatment until it reached 3.8 mN / dtex and the maximum carbonization temperature being 1,600°C, the procedure was the same as in Example 2. The crystallite size Lc was 2.4 nm and the strand tensile modulus E was 300 GPa. Furthermore, the number of fluff particles in the carbon fiber bundle was 2 particles / m, and 3% of the fluff particles had a double-layered cross-section, indicating good quality when the carbon fiber bundle was unwound. The obtained evaluation results are shown in Tables 1 and 2.

[0117] (Example 9) Except for using a copolymer composed of acrylonitrile, itaconic acid, and n-butyl acrylate, and setting the single fiber fineness of the polyacrylonitrile-based precursor fiber bundle to 2.2 dtex, the procedure was carried out in the same manner as in Example 2, resulting in an average single fiber diameter B of 10.5 μm for the carbon fiber bundle. Furthermore, the number of fuzz particles in the carbon fiber bundle was 35 particles / m, and 6% of the fuzz particles had a double-layered cross-section, indicating good quality when the carbon fiber bundle was unwound. The obtained evaluation results are shown in Tables 1 and 2.

[0118] (Example 10) Density 1.38~1.50g / cm 3 Except for setting the heat treatment temperature to 285°C, the procedure was the same as in Example 4. The number of fluff particles in the carbon fiber bundle was 16 particles / m, and 2% of the fluff particles had a double-layered cross-section. The quality of the carbon fiber bundle when unwound was very good. The obtained evaluation results are shown in Tables 1 and 2.

[0119] (Example 11) Except for setting the final carbonization temperature to 1,450°C, the procedure was the same as in Example 4. The number of fluff particles in the carbon fiber bundle was 17 particles / m, and 3% of the fluff particles had a double-layered cross-section. The quality of the carbon fiber bundle at the time of winding was very good. The obtained evaluation results are shown in Tables 1 and 2.

[0120] (Example 12) Yarn width W: 9mm, Density: 1.32~1.35g / cm 3 The heat treatment temperature until this is achieved is 260°C, and the density is 1.38-1.50 g / cm³. 3 The heat treatment temperature until this is achieved is 279°C, and the density is 1.38-1.50 g / cm³. 3 Except for setting the tension of the flame-resistant fiber bundle to 1.7 mN / dtex during heat treatment until it reached the desired state, the procedure was the same as in Example 4. The number of fluff particles in the carbon fiber bundle was 2 particles / m, and 1% of the fluff particles had a double-layered cross-section. The quality of the carbon fiber bundle when unwound was very good. The obtained evaluation results are shown in Tables 1 and 2.

[0121] (Comparative Example 1) Except for setting the yarn width W to 4 mm, the procedure was the same as in Example 1. The number of fluff particles in the carbon fiber bundle was 42 particles / m, and 27% of the fluff particles had a double-layered cross-section. This resulted in a high amount of fluff during winding of the carbon fiber bundle, leading to a deterioration in quality. The evaluation results obtained are shown in Tables 1 and 2.

[0122] (Comparative Example 2) Except for setting the yarn width W to 10 mm, the procedure was the same as in Example 2. The number of fluff particles in the carbon fiber bundle was 4 particles / m, and 0% of the fluff particles had a double-layered cross-section. However, the strand tensile strength decreased to 5.0 GPa, which caused friction near the roller during winding of the carbon fiber bundle, resulting in more fluff and a deterioration in quality. The obtained evaluation results are shown in Tables 1 and 2.

[0123] (Comparative Example 3) Except for increasing the number of filaments N to 51,000, the same procedure as in Example 2 was followed. The number of fluff particles in the carbon fiber bundle was 260 particles / m, and 70% of the fluff particles had a double-layered cross-section. This resulted in a high fluff count and poor quality when unwinding the carbon fiber bundle. The evaluation results are shown in Tables 1 and 2.

[0124] (Comparative Example 4) Except for changing the number of filaments N to 3,000 and the thread width W to 3 mm, the same procedure as in Example 2 was followed. The number of fluff particles in the carbon fiber bundle was 3 particles / m, and 0% of the fluff particles had a double-layered cross-section. However, the strand tensile strength decreased to 5.1 GPa, which caused friction near the roller during winding of the carbon fiber bundle, resulting in more fluff and a deterioration in quality. The obtained evaluation results are shown in Tables 1 and 2.

[0125] (Comparative Example 5) Yarn width W: 5mm, Density: 1.22~1.24g / cm 3Except for setting the heat treatment temperature to 250°C, the procedure was the same as in Example 1. The number of fluff particles in the carbon fiber bundle was 130 particles / m, and 30% of the fluff particles had a double-layered cross-section. When the carbon fiber bundle was unwound, there was a large amount of fluff, resulting in a deterioration of quality. The evaluation results obtained are shown in Tables 1 and 2.

[0126] (Comparative Example 6) Density 1.22~1.24g / cm 3 The heat treatment temperature until this is achieved is 220°C, and the density is 1.38-1.50 g / cm³. 3 Except for setting the heat treatment temperature to 250°C, the procedure was the same as in Example 1. The number of fluff particles in the carbon fiber bundle was 5 particles / m, and 0% of the fluff particles had a double-layered cross-section. However, the strand tensile strength decreased to 5.2 GPa, which caused friction near the roller during unwinding of the carbon fiber bundle, resulting in more fluff and a deterioration in quality. The obtained evaluation results are shown in Tables 1 and 2.

[0127] (Comparative Example 7) Density 1.22~1.24g / cm 3 The heat treatment temperature until this is achieved is 230°C, and the density is 1.38-1.50 g / cm³. 3 The heat treatment temperature was set to 275°C, but the final density of the flame-resistant fiber bundle was 1.36 g / cm³. 3 The results were as follows. Otherwise, the process was the same as in Example 1, and the number of fluff particles in the carbon fiber bundle was 64 particles / m, indicating a high amount of fluff during the winding of the carbon fiber bundle and a deterioration in quality. The obtained evaluation results are shown in Tables 1 and 2.

[0128] (Comparative Example 8) Except for setting the single fiber fineness of the polyacrylonitrile-based precursor fiber bundle to 0.7 dtex, the same procedure as in Comparative Example 8 was followed. This resulted in an average single fiber diameter B of 5.5 μm in the carbon fiber bundle, and increased fuzzing due to friction. Consequently, there was a lot of fuzz when unwinding the carbon fiber bundle, resulting in a deterioration of quality. The evaluation results obtained are shown in Tables 1 and 2.

[0129] (Comparative Example 9) Following Example 2 of Japanese Patent Publication No. 2007-314901, the number of filaments was set to 24,000, flame retardation was performed at 240°C for 130 minutes, and the final carbonization temperature was set to 1,450°C. Except for these other factors, the process was the same as in Example 1, resulting in a final flame-retardant fiber bundle density of 1.35 g / cm³. 3 As a result, the number of fluff particles in the carbon fiber bundle was 48 / m, and there was a large amount of fluff when the carbon fiber bundle was unwound, resulting in a deterioration of quality. The evaluation results obtained are shown in Tables 1 and 2.

[0130] (Comparative Example 10) Following Example 1 of JP 2018-145541, the copolymer was made from acrylonitrile and 2-hydroxyethyl methacrylate, the single fiber fineness of the polyacrylonitrile precursor fiber bundle was set to 4.0 dtex, and the number of filaments was set to 3,000. Flame-retardant treatment was performed under the conditions shown in Table 2, and the density of the flame-retardant fiber bundle was 1.39 mg / m², similar to JP 2018-145541. 3 As a result, when the other conditions were the same as in Example 1, the average single fiber diameter B of the carbon fiber bundle became 13.1 μm. Consequently, the double structure of the carbon fiber bundle deteriorated to 82%, the number of fluff particles in the carbon fiber bundle was 60 particles / m, and 35% of the fluff particles in the carbon fiber bundle had a double structure in cross-section. This resulted in a large amount of fluff during winding of the carbon fiber bundle, leading to a deterioration in quality. The obtained evaluation results are shown in Tables 1 and 2.

[0131] (Comparative Example 11) Except for setting the single fiber fineness of the polyacrylonitrile precursor fiber bundle to 3.0 dtex, the procedure was the same as in Example 9. As a result, the average single fiber diameter B of the carbon fiber bundle became 12.0 μm, and the strand tensile modulus E decreased to 213 GPa. Furthermore, the number of fluff particles in the carbon fiber bundle was 110 particles / m, and 45% of the fluff particles had a double-layered cross-section. This resulted in a high fluff count and degraded quality when unwinding the carbon fiber bundle. The obtained evaluation results are shown in Tables 1 and 2.

[0132] (Comparative Example 12) Except for setting the maximum carbonization temperature to 1,150°C, the procedure was the same as in Example 2. This resulted in a crystallite size Lc of 1.4 nm and a decrease in the strand tensile modulus E to 215 GPa. Furthermore, the number of fluff particles in the carbon fiber bundle was 4 particles / m, and 4% of the fluff particles had a double-layered cross-section. However, friction occurred near the roller during the unwinding of the carbon fiber bundle, resulting in a large amount of fluff and a deterioration in quality. The obtained evaluation results are shown in Tables 1 and 2.

[0133] (Comparative Example 13) In the flame-retardant treatment process, the density is 1.38 to 1.50 g / cm³. 3 Except for setting the tension of the flame-resistant fiber bundle to 1.2 mN / dtex, the procedure was the same as in Example 2, resulting in a strand tensile modulus of 225 GPa. Furthermore, the number of fluff particles in the carbon fiber bundle was 6 particles / m, and 4% of the fluff particles had a double-layered cross-section. However, friction occurred near the roller during the unwinding of the carbon fiber bundle, resulting in a large amount of fluff and a deterioration in quality. The obtained evaluation results are shown in Tables 1 and 2.

[0134] (Comparative Example 14) In the flame-retardant treatment process, the density is 1.38 to 1.50 g / cm³. 3 Except for setting the tension of the flame-resistant fiber bundle to 4.5 mN / dtex until it reached the desired state, the procedure was the same as in Example 2. However, the number of fuzz particles in the carbon fiber bundle was 80 particles / m, resulting in a large amount of fuzz during winding of the carbon fiber bundle and a deterioration in quality. The obtained evaluation results are shown in Tables 1 and 2.

[0135] (Comparative Example 15) Except for setting the maximum carbonization temperature to 2,100°C, the procedure was the same as in Example 2. The crystallite size Lc became 2.9 nm and the strand tensile modulus E increased to 320 GPa. However, the number of fuzz particles in the carbon fiber bundle was 45 particles / m, indicating that friction occurred near the roller during the unwinding of the carbon fiber bundle, resulting in a high fuzz count and degraded quality. The obtained evaluation results are shown in Tables 1 and 2.

[0136] (Comparative Example 16) Density 1.38~1.50g / cm 3 The procedure was carried out in the same manner as in Example 2, except that the heat treatment temperature until the desired result was 305°C. The resulting density was 1.38-1.50 g / cm³. 3 During the process, the flame-resistant fiber bundles broke, and neither flame-resistant fiber bundles nor carbon fiber bundles were obtained. The evaluation results obtained are shown in Tables 1 and 2.

[0137] [Table 1]

[0138] [Table 2]

Claims

1. A carbon fiber bundle having an average single fiber diameter B of 6.9 to 11.0 μm, a resin-impregnated strand tensile modulus E of 230 to 310 GPa, 40 or fewer fluffs per meter within the carbon fiber bundle, and 1 to 25% of the fluffs within the carbon fiber bundle having a double cross-sectional structure.

2. The carbon fiber bundle according to claim 1, wherein the proportion of fluff within the carbon fiber bundle in which the area ratio of the cross-section perpendicular to the fiber axis is 50% or less is 0 to 3%.

3. The carbon fiber bundle according to claim 1 or 2, wherein the tensile modulus E of the resin-impregnated strand and the crystallite size Lc (nm) satisfy the relationship given by formula (1). 50 × Lc + 130 ≤ E ≤ 50 × Lc + 180 ... (1)

4. A carbon fiber bundle according to claim 1 or 2, wherein the crystallite size Lc (nm) is 1.5 to 2.5 nm.

5. A carbon fiber bundle according to claim 1 or 2, wherein the thread width W is 5 to 8 mm.

6. A carbon fiber bundle according to claim 1 or 2, wherein the number of filaments N is 10,000 to 50,000.

7. The carbon fiber bundle according to claim 1 or 2, wherein the knot strength A (MPa) and the average single fiber diameter B (μm) satisfy the relationship given by formula (2). -88B+1360≦A...(2)

8. The carbon fiber bundle according to claim 1 or 2, wherein the resin-impregnated strand has a tensile strength of 5.5 to 7.0 GPa.

9. The carbon fiber bundle according to claim 1 or 2, wherein the area ratio of the outer layer to the entire cross-section perpendicular to the fiber axis of a single fiber is 85 to 95 area %.

10. In a process of heat-treating a polyacrylonitrile-based precursor fiber bundle with a single fiber fineness of 0.9 to 2.2 dtex under an oxidizing atmosphere at 200 to 300°C, when the heat generation rate of the single fiber is q (J / g / s), the number of filaments N is N (threads), the single fiber fineness of the flame-resistant fiber bundle is d (dtex), and the yarn width is W (mm), the density is 1.22 to 1.24 g / cm³. ３ The heat generation rate Q, calculated using equation (3), is 150 to 500 J / m² until it reaches this point. ２ After heat treatment to achieve a density of 1.38–1.50 g / cm³, the density is 1.38–1.50 g / cm³. ３ A method for producing carbon fiber bundles, comprising: heat-treating the fiber bundles to obtain flame-resistant fiber bundles by setting the tension of the fiber bundles to 1.6 to 4.0 mN / dtex until a flame-resistant fiber bundle is obtained; and then heat-treating the flame-resistant fiber bundles in an inert atmosphere at 1,200 to 1,600°C to obtain carbon fiber bundles. Q=q×N×d / W / 10...(3)

11. In the step of heat-treating under the oxidizing atmosphere, the heat generation rate Q obtained by the formula (3) is 150 to 500 J / m ３ until the density reaches 1.22 to 1.24 g / cm ２ After heat-treating so that the heat generation rate Q obtained by the formula (3) is 300 to 1200 J / m ３ until the density reaches 1.32 to 1.35 g / cm ２ Then, heat-treat so that the heat generation rate Q obtained by the formula (3) is 900 to 1500 J / m ３ until the density reaches 1.38 to 1.50 g / cm ２ The method for producing a carbon fiber bundle according to claim 10, wherein the heat treatment is performed so that the heat generation rate Q obtained by the formula (3) is 900 to 1500 J / m

Images