Silicon nitride substrate

Abstract

A silicon nitride substrate according to the present invention has a hardness distribution obtained by nanoindentation testing under the condition of a load of 0.5 mN. The ratio R1 (HA10 / HA90) of the cumulative 10% hardness HA10 in the region of hardness less than 15.00 GPa to the cumulative 90% hardness HA90 in the region of hardness less than 15.00 GPa is 0.25 or more. The ratio R2 (HS10 / HS90) of the cumulative 10% hardness HS10 in the region of hardness 15.00 GPa or more to the cumulative 90% hardness HS90 in the region of hardness 15.00 GPa or more is 0.52 or more.　The present invention can provide a silicon nitride substrate having small temperature dependence in thermal conductivity and dielectric loss, high reliability, and high uniformity at the microstructure level.

Description

Silicon nitride substrate

[0001] The present invention relates to silicon nitride substrates.

[0002] Silicon nitride has excellent mechanical strength, thermal conductivity, and electrical insulation properties, making it suitable for use in semiconductor module substrates, etc. High uniformity is essential for the substrate to exhibit high reliability.

[0003] Regarding the uniformity of silicon nitride substrates, for example, Patent Document 1 proposes a silicon nitride substrate with a small difference in composition between the center and edge portions of the substrate surface. Specifically, the magnesium, yttrium, and oxygen amounts are measured at the center and edge portions of the substrate surface, and the silicon nitride substrate has a small difference between the center and edge portions. It describes that by achieving an extremely high degree of uniformity in the distribution of auxiliary agents from a macroscopic perspective, warping of the silicon nitride substrate can be stably suppressed.

[0004] WO2020 / 203787

[0005] As in Patent Document 1, it is possible to control warpage by controlling uniformity from a macroscopic perspective and minimizing the difference in the amount of additive and oxygen between the center and the edges of the silicon nitride substrate. However, since thermal conductivity and electrical properties are physical properties that depend on the silicon nitride particles and additive phases, in order to achieve high reliability in these physical properties, it is necessary to consider the uniformity of the silicon nitride particles, additive phases, and other properties from a microscopic perspective at the microstructural level. Therefore, an object of the present invention is to provide a silicon nitride substrate that has high uniformity at the microstructural level.

[0006] In order to solve the above problems, the inventors conducted extensive research. As a result, they discovered that by controlling the raw material silicon nitride powder and firing conditions, it is possible to obtain a silicon nitride substrate with high uniformity at the microstructural level. They also confirmed that the uniformity at the microstructural level can be evaluated by nanoindentation testing, which led to the completion of the present invention.

[0007] That is, the present invention provides a silicon nitride substrate in which, in a hardness distribution obtained by a nanoindentation test under a load of 0.5 mN, the ratio R1 (HA10 / HA90) of the cumulative 10% hardness HA10 in the hardness range of less than 15.00 GPa to the cumulative 90% hardness HA90 in the hardness range of less than 15.00 GPa is 0.25 or more, and the ratio R2 (HS10 / HS90) of the cumulative 10% hardness HS10 in the hardness range of 15.00 GPa or more to the cumulative 90% hardness HS90 in the hardness range of 15.00 GPa or more is 0.52 or more.

[0008] The silicon nitride substrate of the present invention makes it possible to reduce the temperature dependence of thermal conductivity and dielectric loss, and to provide a silicon nitride substrate with excellent reliability.

[0009] Fig. 1 is a cross-sectional view of a firing container used in Example 1 of the present invention; Fig. 2 is a cross-sectional view of a firing container used in Example 2 of the present invention; Fig. 3 is a cross-sectional view of a firing container used in Comparative Example 1 of the present invention (a firing container that has been widely used in the past).

[0010] In the silicon nitride substrate of the present invention, in a hardness distribution obtained by a nanoindentation test under a load of 0.5 mN, the ratio R1 (HA10 / HA90) of the cumulative 10% hardness HA10 in the hardness range of less than 15.00 GPa to the cumulative 90% hardness HA90 in the hardness range of less than 15.00 GPa is 0.25 or more, and the ratio R2 (HS10 / HS90) of the cumulative 10% hardness HS10 in the hardness range of 15.00 GPa or more to the cumulative 90% hardness HS90 in the hardness range of 15.00 GPa or more is 0.52 or more.

[0011] Nanoindentation testing uses a tiny indenter to measure indentation hardness, allowing for evaluation of microscopic regions. Typically, silicon nitride substrates contain silicon nitride particles and an additive phase derived from a sintering additive. If the silicon nitride particles and additive phase are homogeneous, it can be said that the substrate has high uniformity at the microstructure level. Note that nanoindentation testing must be performed with a surface roughness of 0.2 μm or less to prevent the effects of surface irregularities. With a surface roughness of 0.2 μm or less, evaluation can be performed with almost no effect from surface irregularities. The surface roughness of the specimen can be reduced to 0.2 μm or less by mirror polishing.

[0012] Through research by the present inventors, it was found that the uniformity of the silicon nitride particles and the auxiliary phase can be evaluated by nanoindentation testing. That is, by performing nanoindentation testing under a load of 0.5 mN, a hardness distribution can be obtained in a microscopic region. The hardness reflects the properties of the silicon nitride particles or the auxiliary phase at each measurement point, and a sharp hardness distribution indicates high uniformity. Generally, silicon nitride particles and the auxiliary phase have different hardnesses, with silicon nitride particles exhibiting a higher hardness. In nanoindentation testing under the above conditions, it was believed that the properties of the silicon nitride particles can be evaluated by the hardness range of 15.00 GPa or more, and the properties of the auxiliary phase can be evaluated by the hardness range of less than 15.00 GPa. Furthermore, if the hardness distribution is sharp, i.e., the ratio of the cumulative 10% hardness to the cumulative 90% hardness (the value obtained by dividing the cumulative 10% hardness by the cumulative 90% hardness) is large, it can be said that the uniformity is high.

[0013] In the present invention, the term "cumulative 10% hardness" refers to the hardness that is the 10% cumulative hardness from the smallest when the hardness measurement data obtained within the hardness range, i.e., hardnesses of 15.00 GPa or more and less than 15.00 GPa, are sorted in ascending order of hardness. Similarly, the term "cumulative 90% hardness" refers to the 90% cumulative hardness from the smallest when the measurement data obtained are sorted in ascending order of hardness. Specifically, when n pieces of measurement data are obtained, the 0.1nth hardness value in the ascending order of hardness is the 10% cumulative hardness, and the 0.9nth hardness value is the 90% cumulative hardness. If 0.1n and 0.9n are not integers, the cumulative 10% hardness and the cumulative 90% hardness can be calculated by rounding them to integers.

[0014] As described above, the properties of the auxiliary phase can be evaluated by the hardness region below 15.00 GPa. Therefore, in the hardness distribution obtained by the nanoindentation test, the ratio R1 (HA10 / HA90) of the cumulative 10% hardness (HA10) in the region below 15.00 GPa to the cumulative 90% hardness (HA90) in the region below 15.00 GPa is calculated. In the silicon nitride substrate of the present invention, R1 is 0.25 or more, preferably 0.26 or more, and more preferably 0.27 or more. The upper limit of R1 is not particularly limited and may be any value below 1.00, which is the theoretical maximum value, but is usually 0.90 or less, particularly 0.75 or less, or even 0.50 or less.

[0015] Furthermore, since the properties of silicon nitride particles can be evaluated based on the hardness range of 15.00 GPa or more, the hardness distribution obtained by the nanoindentation test is evaluated by calculating the ratio R2 (HS10 / HS90) of the cumulative 10% hardness (HS10) in the hardness range of 15.00 GPa or more to the cumulative 90% hardness (HS90) in the hardness range of 15.00 GPa or more. The silicon nitride substrate of the present invention has an R2 of 0.52 or more. There is no particular upper limit for R2, and it may be any value up to 1.00, which is the theoretical maximum value, but it is usually 0.90 or less, particularly 0.75 or less.

[0016] When R1 and R2 are within the above ranges, the silicon nitride substrate has high uniformity at the microstructure level for both the silicon nitride particles and the auxiliary phase. The silicon nitride substrate exhibits reduced temperature dependence of thermal conductivity and dielectric loss. While the reason for this is unclear, the inventors believe it is as follows.

[0017] Thermal conductivity is strongly influenced by the internal structure of a silicon nitride substrate, specifically the microstructural structure of the silicon nitride particles and additive phases that make up the silicon nitride substrate. Therefore, ensuring sufficient heat transfer paths is important. Therefore, significant changes in the internal structure of a silicon nitride substrate can alter the heat transfer paths and significantly reduce thermal conductivity. Temperature changes also cause changes in the internal structure of a silicon nitride substrate due to factors such as expansion and contraction. If the internal structure of a silicon nitride substrate is uniform at the microscopic level, these changes are likely to occur uniformly and significant damage to the heat transfer paths is unlikely. However, if the internal structure of a silicon nitride substrate is not uniform at the microscopic level, significant localized changes in the internal structure can easily occur, causing significant damage to the heat transfer paths. As a result, it is inferred that the temperature dependence of thermal conductivity is greater when the internal structure of a silicon nitride substrate is not uniform at the microscopic level compared to when the internal structure is uniform at the microscopic level. Dielectric loss is also strongly affected by the internal structure of the silicon nitride substrate, and it is presumed that when changes in the internal structure due to temperature change are small, that is, when the internal structure of the silicon nitride substrate is uniform at a microscopic level, the temperature dependence of dielectric loss will be small.

[0018] As described above, the silicon nitride substrate of the present invention comprises silicon nitride particles and an auxiliary phase. The auxiliary phase is composed of components derived from a sintering aid. The blending ratio of the silicon nitride particles to the auxiliary phase is not particularly limited, but if the auxiliary phase is too little, sintering may be insufficient, resulting in insufficient mechanical properties such as strength, while if the auxiliary phase is too much, the advantages of the silicon nitride substrate may not be fully realized. Therefore, in the hardness distribution obtained by the nanoindentation test, the number of points having a hardness of 15.00 GPa or more is preferably 60% or more, more preferably 70% or more, and preferably 95% or less, more preferably 90% or less.

[0019] The silicon nitride substrate of the present invention preferably has a Young's modulus distribution obtained by a nanoindentation test under a load of 0.5 mN, in which the ratio R1 (EA10 / EA90) of the cumulative 10% Young's modulus EA10 in the region where the Young's modulus is less than 225 GPa to the cumulative 90% Young's modulus EA90 in the region where the Young's modulus is less than 225 GPa is 0.46 or more, and the ratio R2 (ES10 / ES90) of the cumulative 10% Young's modulus ES10 in the region where the Young's modulus is 225 GPa or more to the cumulative 90% Young's modulus ES90 in the region where the Young's modulus is 225 GPa or more is 0.71 or more. By ensuring that the above-mentioned ratios R1 (EA10 / EA90) and R2 (ES10 / ES90) are at least certain levels, it becomes possible to suppress the occurrence of microcracks when cutting the silicon nitride substrate, making it possible to provide a highly reliable silicon nitride substrate that is less likely to break due to mechanical or thermal loads during use.

[0020] By performing a nanoindentation test under a load of 0.5 mN, the Young's modulus distribution of a micro-region can be obtained. The Young's modulus reflects the properties of the silicon nitride particles or the auxiliary phase at each measurement point, and a sharp Young's modulus distribution indicates high uniformity. Generally, silicon nitride particles and the auxiliary phase have different Young's moduli, with silicon nitride particles exhibiting a higher Young's modulus. In a nanoindentation test under the above conditions, it was considered that the properties of the silicon nitride particles could be evaluated by the Young's modulus range of 225 GPa or more, and the properties of the auxiliary phase could be evaluated by the Young's modulus range of less than 225 GPa. Furthermore, if the Young's modulus distribution is sharp, that is, if the ratio of the Young's modulus at 10% cumulative to the Young's modulus at 90% cumulative (the value obtained by dividing the Young's modulus at 10% cumulative by the Young's modulus at 90% cumulative) is large, it can be said that the uniformity is high.

[0021] In the present invention, the term "10% cumulative Young's modulus" refers to the Young's modulus at 10% cumulative from the smallest value when the measurement data for Young's moduli obtained within the range of Young's moduli, i.e., at Young's moduli of 225 GPa or more and less than 225 GPa, are sorted in ascending order of Young's modulus. Similarly, the term "90% cumulative Young's modulus" refers to the Young's modulus at 90% cumulative from the smallest value when the measurement data obtained are sorted in ascending order of Young's modulus. Specifically, when n pieces of measurement data are obtained, the Young's modulus is sorted in ascending order of Young's modulus, and the value of the 0.1nth Young's modulus is the 10% cumulative Young's modulus, and the value of the 0.9nth Young's modulus is the 90% cumulative Young's modulus. If 0.1n and 0.9n are not integers, the cumulative 10% Young's modulus and the cumulative 90% Young's modulus can be calculated by rounding them to integers.

[0022] As mentioned above, the properties of the auxiliary phase can be evaluated based on the Young's modulus region below 225 GPa. Therefore, the Young's modulus distribution obtained by the nanoindentation test is evaluated by calculating the ratio R1 (EA10 / EA90) of the cumulative 10% Young's modulus (EA10) in the region below 225 GPa to the cumulative 90% Young's modulus (EA90) in the region below 225 GPa. For the silicon nitride substrate of the present invention, R1 (EA10 / EA90) is preferably 0.46 or greater, more preferably 0.47 or greater, and even more preferably 0.48 or greater. The upper limit of R1 (EA10 / EA90) is not particularly limited and may be any value below the theoretical maximum value of 1.00, but is typically 0.90 or less, particularly 0.80 or less, and even more preferably 0.60 or less.

[0023] Furthermore, since the properties of silicon nitride particles can be evaluated based on the Young's modulus in the region of 225 GPa or more, the Young's modulus distribution obtained by the nanoindentation test is evaluated by calculating the ratio R2 (ES10 / ES90) of the cumulative 10% Young's modulus (ES10) in the region of 225 GPa or more to the cumulative 90% Young's modulus (ES90) in the region of 225 GPa or more. The silicon nitride substrate of the present invention preferably has an R2 (ES10 / ES90) of 0.71 or more, more preferably 0.72 or more. The upper limit of R2 (ES10 / ES90) is not particularly limited and may be any value up to the theoretical maximum value of 1.00, but is typically 0.90 or less, particularly 0.80 or less.

[0024] When R1 (EA10 / EA90) and R2 (ES10 / ES90) are within the above ranges, the silicon nitride substrate has high uniformity at the microstructure level for both the silicon nitride particles and the auxiliary phase. This silicon nitride substrate can suppress the occurrence of microcracks. While the reason for this is unclear, the inventors believe it to be as follows.

[0025] When a force is applied to cut a silicon nitride substrate, stress is applied to the silicon nitride substrate. This stress causes non-uniform strain within the silicon nitride substrate, resulting in localized large loads and microcracks. Since microcracks occur in small areas, it is believed that non-uniformity at the microstructure level is the cause of microcracks. Here, Young's modulus is an index that indicates the relationship between stress and strain. A uniform Young's modulus at the microstructure level leads to uniform strain, thereby suppressing localized large loads at the microstructure level within the silicon nitride substrate and suppressing microcracks. As mentioned above, silicon nitride substrates contain both silicon nitride particles and an auxiliary phase, and since their Young's moduli are different, achieving complete uniformity is difficult. However, it is believed that increasing the uniformity of the Young's modulus at the microstructure level for the silicon nitride particles and the auxiliary phase reduces non-uniformity in strain generation and suppresses microcrack generation.

[0026] As described above, the silicon nitride substrate has silicon nitride particles and an auxiliary phase. The auxiliary phase is composed of components derived from a sintering aid. The blending ratio of the silicon nitride particles to the auxiliary phase is not particularly limited, but if the auxiliary phase is too little, sintering may be insufficient, resulting in insufficient mechanical properties such as strength, while if the auxiliary phase is too much, the advantages of the silicon nitride substrate may not be fully realized. Therefore, in the Young's modulus distribution obtained by the nanoindentation test, the areas with a Young's modulus of 225 GPa or more are preferably 60% or more, more preferably 70% or more, and preferably 95% or less, more preferably 90% or less.

[0027] The silicon nitride substrate of the present invention has a thermal conductivity (λ ２５℃ The silicon nitride substrate of the present invention has a thermal conductivity (λ ) of preferably 50 W / m·K or more, more preferably 70 W / m·K or more, and even more preferably 90 W / m·K or more at 250°C. ２５０℃ ) and the λ ２５℃ The ratio of (λ ２５０℃ / λ２５℃ ) is preferably 0.60 or more, and more preferably 0.70 or more. ２５０℃ / λ ２５℃ It can be said that the closer λ is to 1.00, the smaller the temperature dependency of thermal conductivity. In general, the thermal conductivity of silicon nitride decreases as the temperature increases. ２５０℃ / λ ２５℃ is 1.00 or less. The thermal conductivity can be measured by a laser flash method.

[0028] The silicon nitride substrate of the present invention has a dielectric loss (ε ２５℃ ) and the dielectric loss at 250 °C (ε ２５０℃ ) and the ratio (ε ２５０℃ / ε ２５℃ ) is preferably 0.60 or more and 1.40 or less, and more preferably 0.70 or more and 1.30 or less. ２５０℃ / ε ２５℃ It can be said that the closer to 1.00 the dielectric loss is to temperature dependency, the smaller it is.

[0029] The silicon nitride substrate of the present invention preferably has a three-point bending strength of 600 MPa or more. Also, the fracture toughness value is 6.0 MPa m １／２ Preferably, the EA10 is 80 GPa or more, and more preferably 90 GPa or more, and the ES90 is 400 GPa or less, and more preferably 380 GPa or less. Furthermore, the silicon nitride substrate of the present invention preferably has an iron content of 20 ppm or less, and the total amount of iron and metal impurities other than iron is preferably 25 ppm or less.

[0030] The silicon nitride substrate of the present invention is preferably plate-shaped. The size of the main surface is not particularly limited, but if the size is too large, it becomes difficult to improve uniformity. Therefore, the main surface is preferably rectangular and is 200 mm x 200 mm or less, and more preferably 100 mm x 100 mm or less. The lower limit of the size of the main surface is preferably 10 mm x 10 mm or more, which allows for easy processing for nanoindentation measurement. Furthermore, the ratio of the long side to the short side of the main surface is not particularly limited. The thickness of the silicon nitride substrate of the present invention is not particularly limited, but is generally 0.1 mm to 2.0 mm.

[0031] The silicon nitride substrate of the present invention can be easily obtained by a method for producing a silicon nitride substrate, which includes at least the following steps: a water vapor heat treatment step in which silicon nitride powder is heated and exposed to water vapor under an oxygen-containing atmosphere; a mixing step in which the silicon nitride powder obtained in the water vapor heat treatment step is mixed with a sintering aid powder to obtain a raw material mixture; a molding step in which a molded body of a desired shape is obtained from the raw material mixture; and a firing step in which the molded body is fired at 1700°C to 1900°C in the presence of 5.0 g or more of an atmosphere adjuster per 100 g of the molded body. The raw material mixture may also be referred to as a molding composition. In order to maintain the above-mentioned R1 (EA10 / EA90) and R2 (ES10 / ES90) values ​​below a certain level in the Young's modulus distribution obtained by nanoindentation testing, the silicon nitride powder preferably has an iron content of 20 ppm or less. From the same perspective, the sintering aid powder preferably has an iron content of 20 ppm or less.

[0032] The steam heat treatment process involves exposing silicon nitride powder to steam under heated and oxygen-containing atmosphere. This process is believed to stabilize and uniform the surface of the silicon nitride powder, thereby producing a silicon nitride substrate with high uniformity at the microstructure level. While the reason for this is unclear, the inventors speculate as follows: Oxygen is required for sintering silicon nitride, and the auxiliary components react with oxygen to form a liquid phase, which then densifies the silicon nitride substrate through liquid-phase sintering. Therefore, variations in the surface oxidation state of the silicon nitride powder are thought to result in variations in the sintering reaction and reduced uniformity at the microstructure level. Oxidation of silicon nitride powder also progresses due to thermal and mechanical stresses during the silicon nitride substrate manufacturing process, such as mixing with sintering aids, molding, and firing. Oxidation due to such stresses during the manufacturing process is difficult to control, and variations in the surface oxidation state are likely to occur. Here, by performing a steam heat treatment prior to firing, the surface oxidation of the silicon nitride powder progresses sufficiently. This makes it difficult for oxygen to reach the silicon nitride portion, making it less likely for oxidation to occur due to the load during the manufacturing process, and it is thought that this reduces the variation in the progress of oxidation due to the load during the manufacturing process, thereby suppressing variation in the sintering reaction and making it possible to obtain a silicon nitride substrate with high uniformity at the microstructure level.

[0033] The steam heating treatment can be carried out by exposing the silicon nitride powder to steam under heating. Specifically, the silicon nitride powder and water are placed in a covered container and heated in an oxygen-containing atmosphere to a temperature at or above the temperature at which steam is generated. The amount of water placed in the sealed container should be equal to or greater than the amount of silicon nitride powder. The heating temperature should be 100°C or higher, at which steam is generated, but from the perspective of reliable and efficient treatment, it is preferably 250°C to 700°C, more preferably 400°C to 600°C. The time for the steam heating treatment is preferably 10 minutes to 200 minutes, more preferably 30 minutes to 100 minutes, from the perspective of reliable and efficient treatment. The atmosphere during the steam heating treatment may be an oxygen-containing atmosphere, or may be air. The covered container need not be completely sealed as long as it is filled with steam. The silicon nitride powder and water may be placed in an open-top container, such as a tray, and then placed in the covered container.

[0034] Furthermore, in the above-described manufacturing method, the silicon nitride powder preferably has an iron content of 20 ppm or less. When iron is present, the amount itself is so small that its effect on the overall mechanical properties of the silicon nitride substrate is negligible. However, it is not negligible when viewed locally and is important for achieving uniformity at the microstructural level as in the present invention. The presence of such impurities increases the Young's modulus locally, making stress concentration more likely to occur, and therefore the iron content is presumed to be important. The iron content of the silicon nitride powder is preferably 15 ppm or less, more preferably 10 ppm or less. Furthermore, the silicon nitride powder preferably contains only a small amount of metal impurities other than iron. The total amount of iron and metal impurities other than iron (hereinafter simply referred to as the "metal impurity content") is preferably 25 ppm or less, more preferably 20 ppm or less, and even more preferably 15 ppm or less. In this invention, "metal impurities" refers to metal elements unintentionally contained during the production of a silicon nitride substrate; metal elements contained in the sintering aid used are not included in the "metal impurities." For example, when yttria is used as a sintering aid, the amount of metal elements other than yttrium is the content of metal impurities in the present invention, and when yttria and magnesia are used, the amount of metal elements other than yttrium and magnesium is the content of metal impurities in the present invention.

[0035] In the present invention, the content of metal impurities in the silicon nitride powder, sintering aid powder and silicon nitride substrate can be measured by fluorescent X-ray analysis.

[0036] The content of metal impurities in silicon nitride powder can be reduced by washing (particularly, acid washing) or magnetic separation of the silicon nitride powder. Acid washing can be performed, for example, by treating the silicon nitride powder with a mixed aqueous solution of hydrochloric acid and hydrofluoric acid at 10°C to 80°C for 0.5 to 5 hours. Commercially available hydrochloric acid with a concentration of 35% can be used, or it can be diluted as needed. Commercially available hydrofluoric acid with a concentration of 55% can be used, or it can be diluted as needed. The mixing ratio of hydrochloric acid to hydrofluoric acid can be, for example, 1:5 to 5:1.

[0037] The silicon nitride powder may be made of α-type silicon nitride, β-type silicon nitride, or a mixture thereof. However, β-type silicon nitride is preferred because β-type powder is less susceptible to grain growth during sintering than α-type powder and is easier to control uniformity. The β-phase ratio of the silicon nitride powder is preferably 80% or more, more preferably 90% or more, and even more preferably 99% or more. The β-phase ratio of the silicon nitride powder refers to the peak intensity ratio of the β-phase to the total of the α-phase and β-phase in the silicon nitride powder [100 × (peak intensity of β-phase) / (peak intensity of α-phase + peak intensity of β-phase)], and is determined by powder X-ray diffraction (XRD) measurement using CuKα radiation. More specifically, see C. P. Gazzara and D. R. Messier: Ceram. Bull. , 56 (1977), 777-780, by calculating the weight ratio of the α phase and the β phase of the silicon nitride powder.

[0038] The particle size of the silicon nitride powder is not particularly limited, but the average particle size D50 is preferably 0.5 to 3.0 μm, and more preferably 1.0 to 2.0 μm. The average particle size D50 can be measured using a laser diffraction / scattering particle size distribution analyzer, and refers to the particle size (D50) at which the cumulative curve of the measured volume-based particle size distribution reaches 50% when a volume frequency distribution curve is obtained with the particle size (μm) on the horizontal axis and the volume frequency on the vertical axis. The specific surface area of ​​the silicon nitride powder is 7 to 20 m. ２ / g, and 12 to 15m ２ In the present invention, the specific surface area refers to a BET specific surface area measured by a BET single-point method using nitrogen gas adsorption.

[0039] The sintering aid powder can be any known powder without any particular limitation, and examples thereof include oxides such as yttria, magnesia, ceria, and calcia, and compounds not containing oxygen such as carbonitride compounds and nitride compounds. Examples of carbonitride compounds include, for example, Y ２ Si ４ N ６ C, Yb ２ Si ４ N６ C, Ce ２ Si ４ N ６ C, MgSi ４ N ６ C. Examples of nitride compounds include MgSiN ２ Examples include:

[0040] For the same reasons as for silicon nitride powder, it is preferable to use a sintering aid powder having an iron content of 20 ppm or less. In addition, the sintering aid powder preferably has a small amount of metal impurities other than iron, and the metal impurity content is preferably 25 ppm or less, more preferably 20 ppm or less, and even more preferably 15 ppm or less.

[0041] A raw material mixture is obtained by mixing the silicon nitride powder and the sintering aid powder. The mixing method is not particularly limited; for example, the components may be weighed out in predetermined amounts and mixed dry, or mixed wet using a dispersion medium. Examples of mixing devices for dry mixing include a dry bead mill and an attritor. Examples of mixing devices for wet mixing include an ultrasonic disperser, a bead mill, a ball mill, a roll mill, a homomixer, an ultramixer, a disperser mixer, a penetration-type high-pressure disperser, a collision-type high-pressure disperser, a porous high-pressure disperser, a trap-type high-pressure disperser, a (collision + penetration) type high-pressure disperser, and an ultra-high-pressure homogenizer. Examples of dispersion media for wet mixing include water, alcohol, toluene, and the like.

[0042] The raw material mixture may contain other components in addition to the silicon nitride powder and the sintering aid powder, such as binders, dispersants, plasticizers, and antifoaming agents in addition to the dispersion medium.

[0043] Examples of binders include polyvinyl alcohol, polyvinyl acetal, polyvinylpyrrolidone, acrylic resins, polyacrylamide, urethane resins, polyesters, polyethers, melamine, epoxy resins, cellulose resins, and starch. These binders can be used alone or in combination of two or more. The amount of the binder can be selected appropriately, and can be, for example, typically selected from the range of 1 to 30 parts by mass, and particularly 10 to 25 parts by mass, per 100 parts by mass of the total amount of the silicon nitride powder and the sintering aid powder.

[0044] Dispersants are used to improve the dispersibility of the silicon nitride powder and the sintering aid powder in the molding composition. Generally, surfactants are suitable. Known surfactants can be used without any restrictions. Specific examples of surfactants suitable for use in the present invention include carboxylated trioxyethylene tridecyl ether, diglycerol monooleate, diglycerol monostearate, carboxylated heptaoxyethylene tridecyl ether, tetraglycerol monooleate, hexaglycerol monooleate, sorbitan laurate, sorbitan oleate, sorbitan trioleate, polyoxyethylene sorbitan monolaurate, polyoxyethylene sorbitan monooleate, and polyoxyethylene sorbitan trioleate. These surfactants may be used alone or in combination. The amount of the dispersant can be selected appropriately; for example, it can be selected from the range of 0.1 to 5 parts by mass per 100 parts by mass of the silicon nitride powder and the sintering aid powder combined. Within this range, the upper limit of the amount of dispersant is preferably 3 parts by mass or less, more preferably 2 parts by mass or less, and even more preferably 1 part by mass or less.

[0045] In the molding step, a molded body having a desired shape is obtained from the raw material mixture obtained in the mixing step. The molding method for obtaining the molded body is not particularly limited, and for example, the raw material mixture may be molded by press molding, a slurry of the raw material mixture containing a dispersion medium may be dried using a spray dryer or the like to obtain granules, or a slurry of the raw material mixture containing a dispersion medium may be molded into a sheet by a doctor blade method or the like. Among these, sheet molding by the doctor blade method is preferred because it can be produced easily.

[0046] If the molded body obtained in the molding step contains organic components such as a dispersion medium, binder, dispersant, plasticizer, or antifoaming agent, drying to remove the dispersion medium or degreasing to remove the binder, etc., may be performed before the firing step to facilitate firing. The drying conditions are not particularly limited, but for example, when the dispersion medium is water, the molded body may be heated to about 30°C to 150°C. The degreasing conditions are not particularly limited, but for example, the molded body may be heated to 450°C to 650°C in air or an inert atmosphere such as nitrogen or argon.

[0047] In the firing step of the present invention, the compact is fired at 1700°C to 1900°C in the presence of 5.0 g or more of an atmosphere adjuster per 100 g of compact. The atmosphere adjuster is used to prevent the evaporation of the auxiliary components, and specifically, an atmosphere adjuster having the same chemical composition as the sintering auxiliary powder, or an oxide of a metal element contained in the sintering auxiliary, is used. For example, when yttria is used as the sintering auxiliary, yttria is used as the atmosphere adjuster, and Y is used as the sintering auxiliary. ２ Si ４ N ６ When C is used, Y is used as an atmosphere adjuster. ２ Si ４ N ６ C or yttria is used.

[0048] When multiple compounds are used as the sintering aid powder, the difference between the abundance ratio of the metal elements contained in the atmosphere adjuster and that of the sintering aid powder may be adjusted to within 20%. In this case, a compound having the same chemical composition as the sintering aid powder or an oxide of the metal element contained in the sintering aid may be mixed and used in a desired abundance ratio, or each compound may be placed in the firing vessel so as to achieve the desired abundance ratio.

[0049] Generally, sintering aids volatilize randomly during the firing process. To improve the uniformity of silicon nitride substrates, it is necessary to control the volatilization of sintering aids. By using an atmosphere adjuster, the atmosphere adjuster volatilizes during firing, filling the firing vessel and suppressing the volatilization of sintering aids, which is thought to improve the uniformity of silicon nitride substrates.

[0050] It is important that the firing step be carried out in the presence of 5.0 g or more of an atmosphere adjuster per 100 g of compact. If the amount of atmosphere adjuster is too small, it is not possible to adequately control the volatilization of the sintering aid, and uniformity at the microstructure level cannot be obtained. The amount of atmosphere adjuster is preferably 5.0 g or more, more preferably 7.0 g or more, per 100 g of compact. There is no particular upper limit to the amount of atmosphere adjuster, but if it is too much, production efficiency will be poor, so it is preferably 15.0 g or less, more preferably 10.0 g or less, per 100 g of compact.

[0051] The atmosphere adjuster may be in the form of a powder or a bulk material, but is preferably in the form of a powder because it volatilizes efficiently and fills the firing vessel easily. When the atmosphere adjuster is in the form of a powder, it is preferable that the particle size be 1.0 μm to 2.0 μm from the viewpoint of efficiently filling the firing vessel.

[0052] In the firing process of the present invention, it is also important to increase the airtightness of the firing vessel. If the firing vessel is not airtight, the volatilized atmosphere control agent will not sufficiently fill the firing vessel, resulting in a decrease in the uniformity of the silicon nitride substrate.

[0053] As the firing container, for example, a firing container having a cross section as shown in FIG. 1 or FIG. 2 can be used. The firing container shown in FIG. 1 has a container portion 1 and a lid portion 2, and has two corners at the connection portion 3 between the container portion 1 and the lid portion 2. The firing container shown in FIG. 2 has a different structure of the connection portion 3 from that shown in FIG. 1, but also has two corners. A firing container that has been widely used in the past (FIG. 3) has only one corner at the connection portion between the container portion and the lid portion. By increasing the number of corners at the connection portion, it becomes more difficult for gas inside the firing container to escape to the outside of the firing container through the connection portion, improving airtightness.

[0054] The baking container may have three or more corners, but because a large number of corners increases the effort required to manufacture and maintain the baking container, the number of corners is preferably five or less, and more preferably three or less. The angle of the corner may be more acute or obtuse than 90°, but from the viewpoint of ensuring airtightness and ease of manufacture and maintenance of the baking container, it is preferably 45° to 135°, and more preferably 80° to 100°. Note that, due to the need for sealing, the shape of the container side and the shape of the lid side at the connection part are the same, so the angle on the container side and the angle on the lid side sum to 360°, but the angle of the corner is the smaller value of the angle on the container body side and the angle on the lid side.

[0055] In the firing container, the length of the connection portion 3 is preferably 2.0 mm or more, more preferably 4.0 mm or more. By lengthening the connection portion, it becomes difficult for gas inside the firing container to escape to the outside of the firing container through the connection portion, making it easier to improve airtightness. There is no particular upper limit to the length of the connection portion 3, but since a longer connection portion makes the firing container larger and more difficult to manufacture and handle, it is usually 5.0 mm or less, particularly 4.5 mm or less. The length of the connection portion is the shortest length from the inside to the outside of the portion where the container portion 1 and lid portion 2 of the firing container meet. As described above, the shape of the container portion side and the shape of the lid portion side of the connection portion are the same, so the length of the connection portion can be measured on the container portion side.

[0056] The volume of the firing vessel is not particularly limited, but in order to make it easier to fill the firing vessel with the volatilized atmosphere adjusting agent, the volume of the firing vessel is set to 5250 m per 100 g of atmosphere adjusting agent. ３ Below, especially 5000m ３ The lower limit of the volume of the firing vessel is not particularly limited, but if it is too small, the internal pressure increases due to the volatilized atmospheric control agent, making it difficult to handle. Therefore, the volume of the firing vessel per 100 g of atmospheric control agent is preferably 2500 m or less. ３ Above, especially 2600m ３ It is preferable that the above is set.

[0057] The firing temperature is 1700°C to 1900°C, preferably 1820°C to 1880°C. The firing time is not particularly limited, but is preferably 1 to 30 hours, particularly 3 to 20 hours. The firing is preferably carried out in an inert atmosphere. The firing pressure may be normal pressure or may be increased pressure.

[0058] Examples will be described below to specifically explain the present invention, but the present invention is not limited to these examples. The measurements of the various items in the examples and comparative examples were performed by the following methods.

[0059] (1) Measurement of Hardness Distribution and Young's Modulus Distribution by Nanoindentation Testing A silicon nitride substrate was cut into 10 mm x 10 mm pieces to obtain test specimens. The main surfaces were then mirror-finished, and the surface roughness of the main surfaces of the test specimens was reduced to 0.2 μm or less. The surface roughness of the test specimens was measured using a non-contact 3D measuring device (Keyence Corporation: VR-5000) at five locations in a 1000 μm x 1000 μm test area of ​​the test specimen, and it was confirmed that Ra was 0.2 μm or less in all test areas. A nanoindentation test was performed on the mirror-finished surface of the test specimens in an air atmosphere, with an indenter pressed at 3,600 equally spaced locations within a 30 μm x 30 μm area under a load of 0.5 mN, to determine the hardness and Young's modulus of the microregions. The nanoindentation test was performed at five points near the center and four corners of the main surface of the test specimen, and the hardness distribution and Young's modulus distribution of the microregions were obtained from the results.

[0060] (2) Evaluation of temperature dependence of thermal conductivity After blasting the surface of the silicon nitride substrate, the surface was coated with Au and carbon, and then the thermal conductivity was measured at 25 ° C., and then heated to 250 ° C. and the thermal conductivity was measured. From the results obtained, the thermal conductivity λ at 25 ° C. was calculated. ２５℃ and thermal conductivity λ at 250 ° C. ２５０℃ The ratio of (λ ２５０℃ / λ ２５℃ ) was calculated and the temperature dependence of thermal conductivity was evaluated. ２５０℃ / λ ２５℃ The closer to 1.00 the value is, the smaller the temperature dependency is. The thermal conductivity of the silicon nitride substrate is calculated by multiplying the thermal diffusivity, density, and specific heat, and was measured by the laser flash method using an Advance Riko Co., Ltd. TC-9000. The specific heat of the silicon nitride substrate was set to 0.68 (J / g K). The density of the silicon nitride substrate was measured using an automatic hydrometer (DMA-220H model, manufactured by Shinko Denshi Co., Ltd.).

[0061] (3) Evaluation of temperature dependence of dielectric loss The silicon nitride substrate was cut into 42 mm x 42 mm test pieces and the dielectric loss was measured using the test pieces. The temperature dependence of the dielectric loss was evaluated from the results obtained. The temperature dependence of the dielectric loss was evaluated by first measuring the dielectric loss at 25°C at 1 MHz, then heating to 250°C and measuring the dielectric loss. From the results obtained, the dielectric loss ε at 25°C was calculated. ２５℃ and the dielectric loss ε at 250 °C ２５０℃ The ratio of (ε ２５０℃ / ε ２５℃ ) was calculated and evaluated. ２５０℃ / ε ２５℃ The closer to 1.00 the value is, the smaller the temperature dependency is. The dielectric loss was measured using an HP4192A LF impedance analyzer (manufactured by Agilent Technologies), an HP-16047C test fixture (manufactured by Agilent Technologies), an electrode dimension: φ24 mm, and an inner diameter of a guard electrode: φ26 mm.

[0062] (4) Confirmation of the Presence or Absence of Microcracks After grooves were formed in a silicon nitride substrate by laser scribing, the substrate was cut along the grooves by applying force by hand. The cut surface was observed under a microscope to evaluate the presence or absence of microcracks.

[0063] The silicon nitride substrate was manufactured using the following raw materials containing silicon nitride powder and a sintering aid.

[0064] <Silicon nitride powder> (Silicon nitride powder X) Silicon nitride powder X produced by the same method as silicon nitride powder A described in the examples of WO2021 / 107021. β-phase conversion rate: 99% Average particle size D50: 1.8 μm

[0065] (Silicon nitride powder X2) Silicon nitride powder X produced in the same manner as silicon nitride powder A described in the examples of WO2021 / 107021 was treated with an acid at 60°C for 2 hours using a 1:1 aqueous solution of 35% hydrochloric acid and 55% hydrofluoric acid, filtered, washed with water, vacuum dried at 200°C, and pulverized in a vibrating ball mill for 4 hours to produce silicon nitride powder X2. β-Proportion: 99% Average particle size D50: 1.8 μm Metal impurity content: 10 ppm or less

[0066] <Sintering aids> Yttria (manufactured by Shin-Etsu Chemical Co., Ltd., metal impurity content: 20 ppm or less) Magnesia (manufactured by Ube Materials, metal impurity content: 20 ppm or less) <Binders and dispersants> Binder: acrylic resin (manufactured by Fujikura Kasei Co., Ltd.) Dispersant: Cerna D735 (manufactured by Chukyo Yushi Co., Ltd.)

[0067] Example 1: A tray containing silicon nitride powder X and a tray containing water twice the amount of silicon nitride powder were placed in a thermostatic chamber, heated to 500°C, and left to stand for 1 hour. This resulted in steam heating treatment to obtain silicon nitride powder Y. Next, 100 parts by weight of silicon nitride powder Y, 5 parts by weight of yttria, 3 parts by weight of magnesia, and 0.5 parts by weight of dispersant were weighed out and mixed in a ball mill using water as a solvent in a resin pot and silicon nitride balls for 24 hours. The water was pre-weighed to achieve a slurry concentration of 60% by weight and added to the resin pot. After mixing, 22 parts by weight of binder was added, and the mixture was further mixed for 12 hours to obtain a molding composition in a slurry form. The molding composition was then degassed using a vacuum defoamer (manufactured by Sayama Riken Co., Ltd.) and the viscosity was adjusted to prepare a coating slurry. The viscosity-adjusted coating slurry was then used to form a sheet using a doctor blade method, and the sheet was dried in air at 100°C to evaporate the solvent, yielding a green sheet with a width of 750 mm and a thickness of 420 μm. The green sheet obtained as described above was then degreased in dry air at 550°C to obtain a degreased green sheet. The degreased green sheet was then cut into 100 mm x 100 mm pieces and placed in a firing chamber. The degreased green sheet was fired at 1880°C for 9 hours in a nitrogen atmosphere under a pressure of 0.9 MPa·G to obtain a silicon nitride substrate. The firing chamber used was the one shown in FIG. 1, with two corners at the connection portion and a length of 4.0 mm. Furthermore, 5 g of yttria and 3 g of magnesia were added as atmospheric adjusters per 100 g of green sheet (molded body). The evaluation results of the resulting silicon nitride substrate are shown in Table 1. In the nanoindentation test, 4915 locations had a hardness of less than 15.00 GPa, and 13085 locations had a hardness of 15.00 GPa or more.

[0068] Example 2 A silicon nitride substrate was obtained in the same manner as in Example 1, except that the firing container used was a container having two corners at the connecting portion, the length of which was 4.0 mm, as shown in Figure 2. The evaluation results of the obtained silicon nitride substrate are shown in Table 1. In the nanoindentation test, 5,730 locations had a hardness of less than 15.00 GPa, and 12,270 locations had a hardness of 15.00 GPa or more.

[0069] Comparative Example 1 A silicon nitride substrate was obtained in the same manner as in Example 1, except that the firing container used was a container having one corner at the connection portion, the length of which was 2.0 mm, as shown in Figure 3. The evaluation results of the obtained silicon nitride substrate are shown in Table 1. In the nanoindentation test, 6,180 locations had a hardness of less than 15.00 GPa, and 11,820 locations had a hardness of 15.00 GPa or more.

[0070] Comparative Example 2 A silicon nitride substrate was obtained in the same manner as in Example 1, except that silicon nitride powder X was used without performing the steam heating treatment. The evaluation results of the obtained silicon nitride substrate are shown in Table 1. In the nanoindentation test, 5,030 locations had a hardness of less than 15.00 GPa, and 12,970 locations had a hardness of 15.00 GPa or more.

[0071]

[0072] Example 3: A tray containing silicon nitride powder X2 and a tray containing twice the amount of water as the silicon nitride powder were placed in a thermostatic chamber, heated to 500°C, and left to stand for 1 hour. This resulted in a steam heating treatment, yielding silicon nitride powder Y2. Next, 100 parts by weight of silicon nitride powder Y2, 5 parts by weight of yttria, 3 parts by weight of magnesia, and 0.5 parts by weight of dispersant were weighed out and mixed in a ball mill using a resin pot and silicon nitride balls with water as the solvent for 24 hours. The water was pre-weighed to achieve a slurry concentration of 60% by weight and added to the resin pot. After mixing, 22 parts by weight of binder was added, and the mixture was further mixed for 12 hours to obtain a molding composition in a slurry form. The molding composition was then degassed using a vacuum defoamer (manufactured by Sayama Riken Co., Ltd.) and the viscosity was adjusted to produce a coating slurry. The viscosity-adjusted coating slurry was then used to form a sheet using the doctor blade method, which was then dried in air at 100°C to evaporate the solvent, yielding a green sheet 750 mm wide and 420 μm thick. The green sheet obtained as described above was then degreased in dry air at 550°C to obtain a degreased green sheet. The degreased green sheet was then cut into 100 mm x 100 mm pieces and placed in a firing chamber. The pieces were fired at 1880°C for 9 hours in a nitrogen atmosphere under a pressure of 0.9 MPa·G to obtain a silicon nitride substrate. The firing chamber used was the one shown in FIG. 1, with two corners at the connection portion and a length of 4.0 mm. Furthermore, 5 g of yttria and 3 g of magnesia were added as atmospheric adjusters per 100 g of green sheet (molded body). The evaluation results of the resulting silicon nitride substrate are shown in Table 2. In the nanoindentation test, 5,165 locations had a hardness of less than 15.00 GPa, 12,835 locations had a hardness of 15.00 GPa or more, 4,915 locations had a Young's modulus of less than 225 GPa, and 13,085 locations had a Young's modulus of 225 GPa or more. The content of metal impurities in the silicon nitride substrate was 25 ppm or less.

[0073] Example 4 A silicon nitride substrate was obtained in the same manner as in Example 3, except that the firing container used was a container having two corners at the connection portion, the length of which was 4.0 mm, as shown in Figure 2. The evaluation results of the obtained silicon nitride substrate are shown in Table 2. In the nanoindentation test, 5,763 locations had a hardness of less than 15.00 GPa, 12,237 locations had a hardness of 15.00 GPa or more, 5,811 locations had a Young's modulus of less than 225 GPa, and 12,189 locations had a Young's modulus of 225 GPa or more. The metal impurity content of the silicon nitride substrate was 25 ppm or less.

[0074] Comparative Example 3 A silicon nitride substrate was obtained in the same manner as in Example 3, except that the firing container used was a container having one corner at the connection portion, the length of which was 2.0 mm, as shown in Figure 3. The evaluation results of the obtained silicon nitride substrate are shown in Table 2. In the nanoindentation test, there were 6,062 locations with a hardness of less than 15.00 GPa, 11,938 locations with a hardness of 15.00 GPa or more, 6,395 locations with a Young's modulus of less than 225 GPa, and 11,605 locations with a Young's modulus of 225 GPa or more.

[0075] Comparative Example 4 A silicon nitride substrate was obtained in the same manner as in Example 3, except that silicon nitride powder X2 was used without performing the steam heating treatment. The evaluation results of the obtained silicon nitride substrate are shown in Table 1. In the nanoindentation test, there were 4,866 locations with a hardness of less than 15.00 GPa, 13,134 locations with a hardness of 15.00 GPa or more, 5,284 locations with a Young's modulus of less than 225 GPa, and 12,716 locations with a Young's modulus of 225 GPa or more.

[0076]

[0077] 1 Container part 2 Lid part 3 Connection part

Claims

1. A silicon nitride substrate, in which, in a hardness distribution obtained by a nanoindentation test under a load of 0.5 mN, the ratio R1 (HA10 / HA90) of the cumulative 10% hardness HA10 in a region where the hardness is less than 15.00 GPa to the cumulative 90% hardness HA90 in the region where the hardness is less than 15.00 GPa is 0.25 or more, and the ratio R2 (HS10 / HS90) of the cumulative 10% hardness HS10 in a region where the hardness is 15.00 GPa or more to the cumulative 90% hardness HS90 in the region where the hardness is 15.00 GPa or more is 0.52 or more.

2. The silicon nitride substrate according to claim 1, wherein in a Young's modulus distribution obtained by a nanoindentation test under a load of 0.5 mN, the ratio R1 (EA10 / EA90) of the cumulative 10% Young's modulus EA10 in the region where the Young's modulus is less than 225 GPa to the cumulative 90% Young's modulus EA90 in the region where the Young's modulus is less than 225 GPa is 0.46 or more, and the ratio R2 (ES10 / ES90) of the cumulative 10% Young's modulus ES10 in the region where the Young's modulus is 225 GPa or more to the cumulative 90% Young's modulus ES90 in the region where the Young's modulus is 225 GPa or more is 0.71 or more.

Images