Heat dissipation member, method for manufacturing heat dissipation member, package, and substrate

Abstract

A heat dissipation member (11) comprises: a sintered material part containing copper and at least one of tungsten and molybdenum; and a plurality of silicon oxide particles dispersed in the sintered material part. In the heat dissipation member (11), with respect to the total weight of copper, tungsten, and molybdenum, the copper content is MCu weight percent, the tungsten content is MW weight percent, the molybdenum content is MMo weight percent, and the silicon oxide content is MSiO2 weight percent in terms of SiO2. 0.9 ≥ MCu / (MCu + MW + MMo) ≥ 0.045 and 0.01 ≥ MSiO2 / (MCu + MW + MMo) ≥ 0.0003 are satisfied.

Description

Heat dissipation member, method of manufacturing heat dissipation member, package and substrate

[0001] The present disclosure relates to a heat dissipation member, a method for manufacturing a heat dissipation member, a package, and a substrate.

[0002] A heat dissipation member is sometimes used to promote heat dissipation from a semiconductor device that operates with a relatively large current, such as a power semiconductor element. The heat dissipation member may be used by joining it to a member made of a material different from the material of the heat dissipation member (hereinafter also referred to as a different material member).

[0003] For example, the package disclosed in Japanese Patent Laid-Open Publication No. 2015-204426 (Patent Document 1) includes a heat sink plate (heat dissipation member) and a ceramic frame (different material member). The heat sink plate is used to dissipate heat generated by electronic components mounted on its upper surface. The ceramic frame is joined to the heat sink plate so as to surround the area where the electronic components are mounted. This joining is performed by brazing. The brazing temperature is approximately 780°C. The ceramic frame is made of, for example, alumina or aluminum nitride.

[0004] The heat sink plate is a metal plate. A metal plate with high thermal conductivity and capable of mitigating package warpage due to differences in linear expansion coefficients during brazing with the ceramic frame is selected. For example, a composite metal plate or a clad metal plate is used. The composite metal plate is formed, for example, by an impregnation method. Specifically, it is formed by impregnating a porous high-melting-point metal plate with copper (Cu). The linear expansion coefficient of high-melting-point metals such as tungsten (W) or molybdenum (Mo) is close to that of ceramic, so the linear expansion coefficient of the heat sink plate can be made close to that of the ceramic frame. Furthermore, Cu has excellent thermal conductivity, which can improve the heat dissipation performance of the heat sink plate.

[0005] When the linear expansion coefficient of the heat sink plate needs to be close to that of the ceramic frame, composite metal plates or clad metal plates are widely used, as mentioned above. However, when matching the linear expansion coefficient is not important, simple metal materials are also widely used, such as pure copper, which can significantly increase thermal conductivity.

[0006] Japanese Patent Application Laid-Open No. 2015-204426

[0007] When a package is subjected to a heat cycle, thermal stress is applied to the joint between the heat dissipation member and the dissimilar material member due to the difference in thermal expansion between them. This can result in damage (typically cracks) to the joint or the dissimilar material member. While the technology disclosed in JP 2015-204426 A can relatively reduce the thermal expansion difference between the heat sink plate (heat dissipation member) and the ceramic frame (dissimilar material member), it is not possible to completely eliminate the thermal expansion difference. Furthermore, because W and Mo have high rigidity, if the W or Mo content of the heat dissipation member is increased to suppress the thermal expansion difference, the Young's modulus of the heat dissipation member also increases, resulting in insufficient relief of thermal stress due to elastic deformation of the heat dissipation member. Therefore, in this case, the thermal expansion difference is more likely to directly lead to damage to the joint or the dissimilar material member. While many materials have a low Young's modulus, a suitable material has not yet been found that not only exhibits excessive thermal expansion but also good thermal conductivity.

[0008] The present disclosure has been made to solve the above-mentioned problems, and one of its purposes is to provide a heat dissipation member that has a thermal expansion coefficient close to that of ceramic materials and good thermal conductivity, while being able to suppress a decrease in reliability of the joint or dissimilar material members due to differences in thermal expansion with the dissimilar material members being joined.

[0009] Aspect 1 is a heat dissipation member comprising a sintered material portion containing copper and at least one of tungsten and molybdenum, and a plurality of silicon oxide particles dispersed in the sintered material portion, wherein the content of copper in the heat dissipation member is M with respect to the total weight of copper, tungsten, and molybdenum. Ｃｕ weight percent, and the tungsten content is MＷ weight percent, and the molybdenum content is M Ｍｏ percent by weight, and the silicon oxide content is SiO ２ Converted to M ＳｉＯ２ weight percent, and 0.9 > M Ｃｕ / (M Ｃｕ +M Ｗ +M Ｍｏ ) ≧ 0.045, and 0.01 ≧ M ＳｉＯ２ / (M Ｃｕ +M Ｗ +M Ｍｏ ) ≧0.0003 is satisfied.

[0010] Aspect 2 is the heat dissipation member according to aspect 1, wherein in the particle size distribution based on the number of particles within the particle size range of 0.2 μm or more and less than 10 μm, the proportion of particles within the particle size range of 0.2 μm or more and less than 1.0 μm is 70% or more.

[0011] Aspect 3 is the heat dissipation member according to aspect 2, wherein each of the plurality of silicon oxide particles has a particle size of less than 10 μm.

[0012] Aspect 4 is the heat dissipation member according to aspect 1, wherein in the particle size distribution based on the number of particles within the particle size range of 0.2 μm or more and less than 5 μm, the proportion of particles within the particle size range of 0.2 μm or more and less than 1.0 μm is 70% or more.

[0013] Aspect 5 is the heat dissipation member according to aspect 4, wherein each of the plurality of silicon oxide particles has a particle size of less than 5 μm.

[0014] Aspect 6 is the heat dissipation member according to aspect 1, wherein in the particle size distribution based on the number of particles within the particle size range of 0.2 μm or more and less than 3 μm, the proportion of particles within the particle size range of 0.2 μm or more and less than 1.0 μm is 70% or more.

[0015] A seventh aspect is the heat dissipation member according to the sixth aspect, wherein each of the plurality of silicon oxide particles has a particle size of less than 3 μm.

[0016] Aspect 8 is the heat dissipation member according to aspect 1, wherein in the particle size distribution based on the number of particles in the particle size range of 0.2 μm or more and less than 2 μm, the proportion of the particle size range of 0.2 μm or more and less than 1.0 μm is 70% or more.

[0017] A ninth aspect is the heat dissipation member according to the eighth aspect, wherein each of the plurality of silicon oxide particles has a particle size of less than 2 μm.

[0018] Aspect 10 is the heat dissipation member according to any one of aspects 1 to 9, wherein 0.80≧M Ｃｕ / (M Ｃｕ +M Ｗ +M Ｍｏ ) ≧0.15 is satisfied.

[0019] Aspect 11 is the heat dissipation member according to any one of aspects 1 to 10, wherein the balance of the heat dissipation member other than copper, tungsten, molybdenum, and silicon oxide is less than 0.5 weight percent of the total weight.

[0020] Aspect 12 is the heat dissipation member according to any one of Aspects 1 to 11, Ｍｏ = 0, and 0.806 ≧ M Ｃｕ / (M Ｃｕ +M Ｗ ) ≧0.075 is satisfied.

[0021] Aspect 13 is the heat dissipation member according to any one of Aspects 1 to 11, Ｗ = 0, and 0.887 ≧ M Ｃｕ / (M Ｃｕ +M Ｍｏ ) ≧0.133 is satisfied.

[0022] Aspect 14 is a method for producing a heat dissipation member according to any one of Aspects 1 to 13, comprising mixing a powder of at least one of tungsten having an average particle size of 0.5 μm to 10 μm and molybdenum having an average particle size of 0.5 μm to 10 μm, copper powder having an average particle size of 1.5 μm to 5.0 μm, and SiO powder having an average particle size of 7 nm to 200 nm. ２and a powder containing copper and tungsten, and molybdenum, to form a mixed powder, wherein the mixed powder contains copper in an amount of M based on the total weight of copper, tungsten, and molybdenum. Ｃｕ（Ｐ） weight percent, and the tungsten content is M Ｗ（Ｐ） weight percent, and the molybdenum content is M Ｍｏ（Ｐ） percent by weight, and the silicon oxide content is SiO ２ Converted to M ＳｉＯ２（Ｐ） weight percent, and 0.9 > M Ｃｕ（Ｐ） / (M Ｃｕ（Ｐ） +M Ｗ（Ｐ） +M Ｍｏ（Ｐ） ) ≧ 0.045, and 0.03 ≧ M ＳｉＯ２（Ｐ） / (M Ｃｕ（Ｐ） +M Ｗ（Ｐ） +M Ｍｏ（Ｐ） )≧0.001 is satisfied, and the manufacturing method further includes a step of heating the mixed powder to a temperature equal to or higher than the melting point of copper.

[0023] Aspect 15 is a method for manufacturing a heat dissipation member according to Aspect 14, further comprising a step of forming at least one green sheet containing the mixed powder and a resin, and the step of firing the mixed powder is carried out by firing the at least one green sheet.

[0024] Aspect 16 is a method for manufacturing a heat dissipation member according to Aspect 15, wherein the at least one green sheet is a plurality of green sheets, and the method further comprises a step of forming a laminate by stacking the plurality of green sheets on each other, and the step of firing the mixed powder is carried out by firing the laminate.

[0025] Aspect 17 is a package comprising a heat dissipation member according to any one of aspects 1 to 13 and a ceramic frame, wherein the heat dissipation member has a heat dissipation surface and a main surface opposite the heat dissipation surface, and the ceramic frame is disposed on the main surface of the heat dissipation member and has an inner surface surrounding a cavity and an outer surface opposite the inner surface.

[0026] Aspect 18 is a substrate comprising a heat dissipation member described in any one of aspects 1 to 13 and a ceramic insulating layer, wherein the heat dissipation member has a heat dissipation surface and a main surface opposite the heat dissipation surface, and the ceramic insulating layer is arranged on the main surface of the heat dissipation member.

[0027] According to the first aspect, the heat dissipation component contains copper, which facilitates good thermal conductivity, and also contains at least one of tungsten and molybdenum, which allows for the adjustment of the thermal expansion coefficient. Furthermore, the inclusion of silicon oxide particles suppresses the Young's modulus, and the inclusion of a moderate amount of silicon oxide prevents significant adverse effects on the thermal conductivity. As a result, the heat dissipation component has moderate thermal expansion and good thermal conductivity, while preventing a decrease in the reliability of the joint or the dissimilar material components due to the difference in thermal expansion between the heat dissipation component and the dissimilar material components to be joined.

[0028] The objects, features, aspects, and advantages of the present disclosure will become more apparent from the following detailed description and the accompanying drawings.

[0029] 5 is a schematic perspective view showing the configuration of a semiconductor module according to a first embodiment, with some parts omitted so that the interior of a cavity can be seen. FIG. 6 is a schematic cross-sectional view of the semiconductor module of FIG. 1 taken along line II-II. FIG. 7 is a schematic cross-sectional view showing the configuration of a package as a component of the semiconductor module of FIG. 2. FIG. 8 is a flow diagram showing an outline of a method for manufacturing a package according to the first embodiment. FIG. 9 is a partial enlarged view of FIG. 5. FIG. 10 is a schematic partial cross-sectional view illustrating a step of the method for manufacturing a package according to the first embodiment. FIG. 11 is a schematic partial cross-sectional view illustrating a step of the method for manufacturing a package according to the first embodiment. FIG. 12 is a schematic cross-sectional view showing the configuration of a semiconductor module according to a second embodiment. FIG. 13 is an electron microscope photograph of a cross-section of a heat dissipation member of Example 2. FIG. 14 is an electron microscope photograph of a cross-section of a heat dissipation member of Example 3. FIG. 15 is an electron microscope photograph of a cross-section of a heat dissipation member of Example 9. FIG. 16 is a graph (a) showing the relationship between the copper content and the stress at 1% strain in a heat dissipation member containing copper and tungsten, and FIG. 17 is a graph (b) showing the relationship between the copper content and the stress at 1% strain in a heat dissipation member containing copper and molybdenum. 17 is a graph (a) showing the relationship between the thermal expansion coefficient and stress at 1% strain in a heat dissipation member containing copper and tungsten, and a graph (b) showing the relationship between the thermal expansion coefficient and stress at 1% strain in a heat dissipation member containing copper and molybdenum. FIG. 18 is a diagram schematically showing the microstructure of a heat dissipation member in which silicon oxide particles are not dispersed. FIG. 19 is a diagram schematically showing the microstructure of a heat dissipation member in which silicon oxide particles are dispersed. FIG. 19 is a schematic cross-sectional view showing the configuration of a semiconductor module according to a fourth embodiment. FIG. 19 is a schematic cross-sectional view showing the configuration of a heat dissipation substrate as a component of the semiconductor module of FIG. 17. FIG. 20 is a schematic partial cross-sectional view illustrating one step of a method for manufacturing a heat dissipation substrate according to the fourth embodiment.

[0030] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0031] First Embodiment (Configuration of Semiconductor Module) FIG. 1 is a schematic perspective view showing the configuration of a semiconductor module 91 according to the first embodiment. FIG. 2 is a schematic cross-sectional view of the semiconductor module 91 taken along line II-II in FIG. 1. The semiconductor module 91 has a package 51 and a semiconductor element 8. The semiconductor module 91 may also have wires 9 as wiring members for the semiconductor element 8. The semiconductor module 91 may also have a lid 80 for sealing the cavity CV. The lid 80 may be attached to the package 51 by an adhesive layer 70. Note that in FIG. 1, the lid 80 and the adhesive layer 70 are partially omitted so that the interior of the cavity CV of the package 51 can be partially seen.

[0032] The semiconductor element 8 is typically a power semiconductor element, and in this case, the semiconductor module 91 is a power module. The power semiconductor element may be for radio frequency (RF) use, and in this case, the semiconductor module 91 is an RF power module. The semiconductor element 8 is not limited to a power semiconductor element, and may be, for example, an LSI (Large-Scale Integration) or IC (Integrated Circuit) that operates at high power. Although one semiconductor element 8 is illustrated in FIGS. 1 and 2, multiple semiconductor elements 8 may be mounted in the package 51. Elements other than the semiconductor element 8, such as passive elements, may also be mounted.

[0033] (Package Configuration) Figure 3 is a schematic cross-sectional view showing the configuration of the package 51 as a component of the semiconductor module 91 (Figure 2). When the package 51 is prepared for manufacturing the semiconductor module 91, as shown in Figure 3, the semiconductor element 8 does not need to be mounted yet. The package 51 has a cavity CV that will be sealed with a lid 80. The package 51 has a heat sink 11 (heat dissipation member) and a ceramic frame 21.

[0034] The heat sink 11 has a heat dissipation surface P1 and a main surface P2 opposite to the heat dissipation surface P1. The heat dissipation surface P1 of the heat sink 11 is typically attached to a support member (not shown). The support member is, for example, a mounting board or a heat dissipation member. The heat sink 11 may have a through-hole (not shown) through which a fastener (e.g., a screw) for attaching the heat sink 11 to the support member passes.

[0035] The ceramic frame 21 is a frame made of ceramic. By using the ceramic frame 21 as the frame of the package 51, it is possible to improve the heat resistance and insulation properties of the package 51. The material of the ceramic frame 21 may contain alumina as a main component, and may also contain a small amount of silica to promote sintering of the ceramic frame 21.

[0036] The ceramic frame 21 is disposed on the main surface P2 of the heat sink 11. The ceramic frame 21 has an inner surface P3 surrounding the cavity CV and an outer surface P4a opposite the inner surface P3. The heat sink 11 may have a side surface P4b that is flush with the outer surface P4a of the ceramic frame 21. The outer edge of the ceramic frame 21 may have a rectangular shape in an in-plane direction perpendicular to the thickness direction, as shown in FIG. 1. The size of each side of the rectangular shape is, for example, 4 mm or more and 40 mm or less. The thickness of the ceramic frame 21 is, for example, 0.1 mm or more and 1 mm or less.

[0037] In the first embodiment, the main surface P2 of the heat sink 11 includes a cavity surface P2a facing the cavity CV and a bonding surface P2b directly bonded to the ceramic frame 21. Thus, the ceramic frame 21 and the heat sink 11 are directly bonded to each other. Here, the expression "directly bonded" means that no components other than those derived from the heat sink 11 and the ceramic frame 21 are detected at the bonded portion.

[0038] The package 51 may have a lead frame 30 (metal terminals). The lead frame 30 is provided on a ceramic frame 21 and is separated from the heat sink 11 by the ceramic frame 21. The lead frame 30 forms an electrical path connecting the inside and outside of the cavity CV. A bonding material (not shown) may be provided between the lead frame 30 and the ceramic frame 21 to bond them to each other. This bonding material may be formed, for example, by Ag sinter bonding. In this case, the bonding material is a mixture of a thermosetting resin (e.g., epoxy resin or silicone resin) and Ag particles. Alternatively, silver solder may be used as the bonding material. In this case, a metallized layer for silver solder is usually formed in advance on the ceramic frame 21.

[0039] In one example of a method for forming the metallization layer, a paste for the metallization layer is printed on a green sheet that will become the ceramic frame 21 before the firing process (described in detail below) for forming the ceramic frame 21 and the heat sink 11. Specifically, a metal powder of at least one of W, Mo, and Cu is first blended with additives, resin, a solvent, and the like, and ceramic powder is further added as needed, followed by kneading to prepare a paste. This paste is then printed, for example, by screen printing, on a green sheet prepared in a previous process. After printing, the green sheet is dried, for example, at a temperature of 110°C for 5 minutes. Alternatively, the metallization layer may be formed by stacking a green sheet containing a metal on the green sheet that will become the ceramic frame 21 before the firing process (described in detail below) for forming the ceramic frame 21 and the heat sink 11.

[0040] The lid 80 (FIGS. 1 and 2) may be made of ceramic, and the ceramic may contain alumina as a main component, for example, substantially alumina. Alternatively, the lid 80 may contain a resin. The resin may be, for example, a liquid crystal polymer. An inorganic filler may be dispersed in the resin, for example, silica particles. Dispersing the inorganic filler in the resin can increase the strength and durability of the lid 80.

[0041] The semiconductor element 8 ( FIG. 2 ) is mounted on the cavity surface P2a ( FIG. 3 ) of the main surface P2 of the heat sink 11 of the package 51. The distance L1 ( FIG. 2 ) between the mounted semiconductor element 8 and the inner surface P3 of the ceramic frame 21 may be 25 μm or less. The distance L1 may even be zero. In other words, the semiconductor element 8 and the inner surface P3 of the ceramic frame 21 may be in contact with each other. Thus, the distance L1 is easier to make smaller than the distance L9 ( FIG. 9 : embodiment 2). This is because the semiconductor element 8 and the brazing filler layer 26 do not interfere with each other. The brazing filler layer 26 has fluidity during its formation and flows inward beyond the inner circumferential surface of the ceramic frame 29 (the surface facing the cavity CV), as shown in FIG. 9 . The portion of the brazing filler layer 26 that flows into the cavity CV forms a fillet 26f at the edge of the cavity CV. The flow distance, i.e., the width of the fillet 26f, tends to be greater than 25 μm. Therefore, to sufficiently reduce the possibility of interference between the fillet 26f and the semiconductor element 8, the distance L9 ( FIG. 9 ) between the semiconductor element 8 and the inner surface of the ceramic frame 29 must be greater than 25 μm. This large gap between the semiconductor element 8 and the ceramic frame 29 results in a smaller mounting area in the cavity CV (the area in which the semiconductor element 8 can be mounted). Furthermore, the length of the wire 9 increases, which typically leads to deterioration of electrical characteristics, such as an unintended increase in inductance.

[0042] The semiconductor element 8 may be mounted using, for example, a solder material (not shown). After mounting the semiconductor element 8, wires 9 ( FIG. 2 ) may be formed to electrically connect the semiconductor element 8 to the lead frame 30. This formation may be performed by wire bonding. Subsequently, the lid 80 may be attached to the package 51. This attachment may be performed using an adhesive layer 70. The adhesive layer 70 may be a thermosetting resin. The adhesive layer 70 is provided on the ceramic frame 21 so as to surround the cavity CV. As shown in FIG. 2 , the adhesive layer 70 may have a portion that is provided on the ceramic frame 21 via the lead frame 30. The thickness of the adhesive layer 70 between the lid 80 and the package 51 is, for example, 100 μm or more and 360 μm or less.

[0043] In the first embodiment, the ceramic frame 21 and the heat sink 11 are formed as a single sintered body SF. Therefore, the ceramic frame 21 and the heat sink 11 are directly bonded to each other. Therefore, no bonding layer (e.g., a brazing filler metal layer 26 ( FIG. 9 : second embodiment)) is provided between the ceramic frame 21 and the heat sink 11 to bond them together. Note that an Ag brazing filler metal is typically used as the brazing filler metal layer 26 ( FIG. 9 : second embodiment). When the brazing filler metal layer 26 contains Ag, Ag migration is likely to occur as indicated by the arrow MG ( FIG. 9 ) if a negative potential is applied to the lead frame 30 for a long period of time relative to the potential of the heat sink 11. This Ag migration can result in insufficient electrical insulation between the heat sink 11 and the lead frame 30. According to the first embodiment, this phenomenon can be prevented.

[0044] (Heat Sink Material) The heat sink 11 is a sintered body including a sintered material portion containing Cu and a high-melting-point metal, and a plurality of silicon oxide particles dispersed in the sintered material portion. The high-melting-point metal has a higher melting point than Cu. The high-melting-point metal used in this embodiment is W and / or Mo, in other words, at least one of W and Mo. The plurality of silicon oxide particles may be sintered together with the sintered material portion. The sintered material portion may constitute a major portion of the heat sink 11.

[0045] In order to improve the heat dissipation performance of the heat sink 11, it is preferable that the material of the heat sink 11 has a high thermal conductivity. Such a high thermal conductivity can be easily obtained by including a sufficient amount of Cu in the heat sink 11. However, since the linear expansion coefficient of Cu is larger than that of typical ceramic materials (e.g., alumina), if the Cu content of the heat sink 11 is excessive, the difference in thermal expansion between the heat sink 11 and the ceramic frame 29 can easily become a problem.

[0046] By including at least one of W and Mo in the heat sink 11 in sufficient proportions, the linear expansion coefficient of the heat sink 11 can be made closer to that of ceramics such as alumina, compared to a heat sink containing almost exclusively Cu. This reduces the thermal expansion difference between the heat sink 11 and the ceramic frame 21. On the other hand, increasing the W or Mo content in the heat sink 11 increases the Young's modulus of the heat sink 11 due to the high rigidity of W and Mo. As a result, the effect of the elastic deformation of the heat sink 11 in mitigating thermal stress is reduced. Therefore, the thermal expansion difference is likely to lead to thermal stress at the joint between the heat sink 11 and the ceramic frame 21 or in the ceramic frame 21 itself. Therefore, the thermal expansion difference is likely to directly damage the joint between the heat sink 11 and the ceramic frame 21 or the ceramic frame 21 itself. Therefore, it is desirable to suppress the Young's modulus of the heat sink 11. For this purpose, the heat sink 11 contains silicon oxide, as will be described in more detail below.

[0047] The heat sink 11 contains Cu, at least one of W and Mo, and silicon oxide. In the heat sink 11, the content of Cu is 0.05% by weight of M with respect to the total weight of Cu, W, and Mo. Ｃｕ wt% (weight percent), and the content of W is M Ｗ wt%, and the Mo content is M Ｍｏ wt%. Therefore, M Ｃｕ wt%+M Ｗ wt%+M Ｍｏ wt% = 100 wt%. Based on this total weight, the content of silicon oxide is SiO ２ Converted to M ＳｉＯ２The specific method for measuring the content of silicon oxide in the heat sink 11 will be described later.

[0048] The composition of the heat sink 11 satisfies the following condition: 0.9 ≧ M Ｃｕ / (M Ｃｕ +M Ｗ +M Ｍｏ ) ≧ 0.045, and 0.01 ≧ M ＳｉＯ２ / (M Ｃｕ +M Ｗ +M Ｍｏ ) ≧ 0.0003 is satisfied. Ｃｕ / (M Ｃｕ +M Ｗ +M Ｍｏ If M is smaller than 0.045, the thermal conductivity of the heat sink will be poor. Ｃｕ / (M Ｃｕ +M Ｗ +M Ｍｏ If M is greater than 0.9, it becomes difficult to match the thermal expansion coefficient with other members such as a ceramic frame or mounting board, which may result in warping or cracks. ＳｉＯ２ / (M Ｃｕ +M Ｗ +M Ｍｏ If M is smaller than 0.0003, the significant effect of suppressing the Young's modulus of the heat sink 11 cannot be obtained. ＳｉＯ２ / (M Ｃｕ +M Ｗ +M Ｍｏ If the value of M is greater than 0.01, the strength of the sintered body cannot be increased to a level sufficient for use as a heat dissipating member. Ｃｕ / (M Ｃｕ +M Ｗ +M Ｍｏ ) ≧ 0.15. Furthermore, the following condition may be satisfied: 0.37 ≧ M Ｃｕ / (M Ｃｕ +M Ｗ +M Ｍｏ ) ≧0.10 may be satisfied.

[0049] The heat sink 11 does not necessarily need to contain Mo, and may satisfy the following conditions: Ｍｏ = 0, and 0.806 ≧ MＣｕ / (M Ｃｕ +M Ｗ ) ≧ 0.075 may be satisfied. Ｃｕ / (M Ｃｕ +M Ｗ If M is smaller than 0.075, the thermal conductivity of the heat sink tends to be insufficient. Ｃｕ / (M Ｃｕ +M Ｗ If the coefficient of thermal expansion of the ceramic substrate is greater than 0.806, it becomes difficult to match the coefficient of thermal expansion of other components such as a ceramic frame or mounting board, which may result in warping or cracks. Ｃｕ / (M Ｃｕ +M Ｗ ) ≧0.10 may be satisfied.

[0050] Alternatively, the heat sink 11 does not necessarily need to contain W, and may satisfy the following conditions: Ｗ = 0, and 0.887 ≧ M Ｃｕ / (M Ｃｕ +M Ｍｏ ) ≧ 0.133 may be satisfied. Ｃｕ / (M Ｃｕ +M Ｍｏ If M is smaller than 0.133, the thermal conductivity of the heat sink tends to be insufficient. Ｃｕ / (M Ｃｕ +M Ｍｏ If M ) is greater than 0.887, it becomes difficult to match the thermal expansion coefficient with other members such as a ceramic frame or mounting board, which may result in warping or cracks. Ｃｕ / (M Ｃｕ +M Ｍｏ ) ≧0.18 may be satisfied.

[0051] In the heat sink 11, the remainder other than Cu, W, Mo, and silicon oxide may be less than 0.5 wt % of the total weight of Cu, W, and Mo. In other words, the heat sink 11 may be substantially composed of Cu, at least one of W and Mo, and silicon oxide.

[0052] Since a plurality of silicon oxide particles are dispersed inside the heat sink 11, the particle size of the silicon oxide particles can be measured by observing the cross section of the heat sink 11 with an electron microscope. A specific method for measuring particle size will be described later. The particle size distribution obtained from this particle size measurement may satisfy at least one of the following first to fourth conditions.

[0053] The first condition is that in the particle size distribution of the plurality of silicon oxide particles in the particle size range of 0.2 μm or more and less than 10 μm, based on the number of particles, the proportion of particles in the particle size range of 0.2 μm or more and less than 1.0 μm (i.e., the proportion of particles in the number of particles) is 70% or more. In this case, preferably, the proportion of particles with a particle size of 10 μm or more among the plurality of silicon oxide particles is 0.1% or less. In other words, the proportion of particles with a particle size of 10 μm or more among the plurality of silicon oxide particles is substantially zero. In other words, each of the plurality of silicon oxide particles has a particle size of less than 10 μm. This further reduces the risk of silicon oxide particles becoming the starting point for damage to the heat sink 11.

[0054] As a second condition, in the particle size distribution of the plurality of silicon oxide particles in the particle size range of 0.2 μm or more and less than 5 μm, based on the number of particles, the proportion of particles in the particle size range of 0.2 μm or more and less than 1.0 μm is 70% or more. In this case, preferably, the proportion of particles with a particle size of 5 μm or more among the plurality of silicon oxide particles is 0.1% or less. In other words, the proportion of particles with a particle size of 5 μm or more among the plurality of silicon oxide particles is substantially zero. In other words, each of the plurality of silicon oxide particles has a particle size of less than 5 μm. This further reduces the risk of silicon oxide particles becoming the starting point for damage to the heat sink 11.

[0055] As a third condition, in the particle size distribution of the plurality of silicon oxide particles in the particle size range of 0.2 μm or more and less than 3 μm, based on the number of particles, the proportion of particles in the particle size range of 0.2 μm or more and less than 1.0 μm is 70% or more. In this case, preferably, the proportion of particles with a particle size of 3 μm or more among the plurality of silicon oxide particles is 0.1% or less. In other words, the proportion of particles with a particle size of 3 μm or more among the plurality of silicon oxide particles is substantially zero. In other words, each of the plurality of silicon oxide particles has a particle size of less than 3 μm. This further reduces the risk of silicon oxide particles becoming the starting point for damage to the heat sink 11.

[0056] As a fourth condition, in the particle size distribution of the plurality of silicon oxide particles in the particle size range of 0.2 μm or more and less than 2 μm, based on the number of particles, the proportion of particles in the particle size range of 0.2 μm or more and less than 1.0 μm is 70% or more. In this case, preferably, the proportion of particles with a particle size of 2 μm or more among the plurality of silicon oxide particles is 0.1% or less. In other words, the proportion of particles with a particle size of 2 μm or more among the plurality of silicon oxide particles is substantially zero. In other words, each of the plurality of silicon oxide particles has a particle size of less than 2 μm. This further reduces the risk of silicon oxide particles becoming the starting point for damage to the heat sink 11.

[0057] (Method for Measuring the Particle Size of Silicon Oxide Particles Dispersed in a Heat Sink) First, the heat sink 11 is cut along the thickness direction. This exposes the cross section of the heat sink 11. Ion milling is performed on a region near the center of the cross section. An image of the region is then captured using an electron microscope. In the experiments described below, a field emission electron probe microanalyzer (FE-EPMA, model: JXA-8500F (manufactured by JEOL Ltd.)) was used as the electron microscope at a measurement magnification of 2000 to 3000 times and an acceleration voltage of 15 kV. As can be seen in Figures 10 to 12, in the images, the white areas represent high-melting-point metals (W and / or Mo), the gray areas represent Cu, and the black areas represent silicon oxide particles. This image data is binarized into black and white using image processing software (ImageJ) to distinguish between regions of silicon oxide particles and other regions. The binarized image data contains multiple regions corresponding to silicon oxide particles. Ellipse fitting is performed for each of these regions. Specifically, for each region, an ellipse with the smallest area that includes the region is determined, and the average of the minor axis and major axis of this ellipse is calculated as the particle size of the silicon oxide particle.

[0058] (Method for measuring the content of silicon oxide in a heat sink) From the image data, the areas of the high-melting-point metal (W and / or Mo), Cu, and silicon oxide particle regions are calculated by black-and-white binarization using image processing software (ImageJ). This area ratio is considered to be the volume ratio of the high-melting-point metal (W and / or Mo), Cu, and silicon oxide particles in the heat sink 11. The weight composition ratio is calculated by dividing this volume ratio by the density of each material. For example, when the density of W is 19.3 g / cm ３ , the density of Mo is 10.2 g / cm ３ , the density of Cu is 8.9 g / cm ３ , silicon oxide (SiO ２ ) density 2.2 g / cm ３ is used in the calculation.

[0059] In addition, when the raw material composition ratio is known, it is considered that using this raw material composition ratio is more accurate than the composition ratio based on the image analysis described above. Therefore, in the experiments described below, the raw material composition ratio is used as the composition ratio between the high-melting-point metal (W and / or Mo) and Cu in the heat sink 11 (heat dissipation member).

[0060] (Package Manufacturing Method) Fig. 4 is a flow diagram that shows a schematic diagram of a method for manufacturing the package 51 (Fig. 3). Figs. 5 to 8 are schematic partial cross-sectional views that illustrate one step of the manufacturing method.

[0061] In step ST6 (FIG. 4), a powder of at least one of W and Mo, a Cu powder, and SiO ２ A mixed powder is formed by mixing the W and Mo powders. The average particle size of the W and Mo powders may be 0.5 μm or more and 10 μm or less, preferably 0.5 μm or more and 3 μm or less, and more preferably 0.5 μm or more and 1.5 μm or less. If the average particle size of the W and Mo powders is smaller than 0.5 μm, the cost of the raw materials increases, and if the average particle size is larger than 10 μm, it becomes difficult to obtain a uniform sintered body due to the difference in specific gravity of the raw materials. The average particle size of the Cu powder may be 1.5 μm or more and 5.0 μm or less. If the average particle size of the Cu powder is smaller than 1.5 μm, the cost of the raw materials increases. If the average particle size is larger than 5.0 μm, it becomes difficult to obtain a uniform sintered body due to the difference in specific gravity of the raw materials. SiO ２ The average particle size of the powder may be 7 nm or more and 200 nm or less, preferably 7 nm or more and 100 nm or less, and more preferably 7 nm or more and 50 nm or less. If the average particle size is larger than 200 nm, it is difficult to obtain a dense sintered body after sintering. On the other hand, SiO ２ Since powder is difficult to manufacture, using this powder as a raw material would significantly increase the raw material cost of the heat sink. ２ The powder may be either crystalline or amorphous, but amorphous SiO ２ Powder is preferred because of its availability.

[0062] SiO ２ The average particle size of the powder is SiO ２ This can be measured using an SEM photograph of the powder. For example, the SEM photograph of SiO ２ Ellipse fitting similar to that described above is performed on each of the multiple particles in the powder, and the average of the minor axis and major axis of the ellipse is taken as the particle size of the particle. For example, the average value of the particle sizes of 100 particles is used as the average particle size. The average particle sizes of W powder and Mo powder may be measured using the Fischer method (Tungsten and Molybdenum Industry Association (Japan) Standard TMIAS0001:1999 Particle Size Test Method). The particle size of Cu powder may be measured using a laser diffraction method, for example, by stirring the powder in isopropyl alcohol (IPA) for 1 minute and then measuring it using a nanoparticle size distribution analyzer SALD-7500nano (manufactured by Shimadzu Corporation).

[0063] In the mixed powder, the copper content is M with respect to the total weight of Cu, Mo, and W. Ｃｕ（Ｐ） % by weight, and the tungsten content is M Ｗ（Ｐ） wt%, and the molybdenum content is M Ｍｏ（Ｐ） wt%, and the silicon oxide content is SiO ２ Converted to M ＳｉＯ２（Ｐ） wt%, and 0.9≧M Ｃｕ（Ｐ） / (M Ｃｕ（Ｐ） +M Ｗ（Ｐ） +M Ｍｏ（Ｐ） ) ≧ 0.045, and 0.03 ≧ M ＳｉＯ２（Ｐ） / (M Ｃｕ（Ｐ） +M Ｗ（Ｐ） +M Ｍｏ（Ｐ） ) ≧0.001 may be satisfied.

[0064] The mixing step is performed by mixing Cu powder and SiO ２ powder to form Cu—SiO ２ a first mixing step for obtaining a mixed powder, and a second mixing step for obtaining a mixed powder of Cu—SiO ２ and a second mixing step of mixing the mixed powder with a powder of at least one of W and Mo. In this case, the mixed powder finally obtained contains SiO ２This increases the proportion of the powder located on the surface of the Cu powder particles. The mixing step may be performed, for example, using a ball mill. After the mixing step, the mixed powder alone may be isolated. However, this isolation is not necessarily required. For example, the mixed powder may form a suspension together with a solvent, a dispersant, a plasticizer, etc. This suspension may then be used as the raw material for the green sheet or molded body described below.

[0065] In step ST11 (FIG. 4), multiple green sheets LG1 to LG9 (see FIG. 6, described below) are formed, which will become the heat sink 11 (FIG. 3) upon firing. In FIG. 6, the structure that will become the heat sink 11 upon firing is a stack of nine green sheets, but the number of green sheets is not particularly limited. The green sheets LG1 to LG9 contain a mixed powder and a resin. The green sheets LG1 to LG9 may be formed by dividing a single green sheet. A known typical method may be used to form the green sheets from the mixed powder. For example, a slurry is first prepared. The slurry is obtained by mixing powders that will become the components of the sintered body with resin, a plasticizer, a solvent, and the like using a ball mill. The slurry is then processed into green sheets using a doctor blade method. The planar shape of the green sheets is determined according to the shape of the desired part. The planar shape of the green sheets used to form the heat sink 11 is typically approximately rectangular.

[0066] In step ST13 (FIG. 4), a frame ceramic green sheet 21G is formed, which is fired to become the ceramic frame 21 (FIG. 3). The powder for forming the frame ceramic green sheet 21G is, for example, Al as the main component. ２ O ３ SiO powder and sintering aid ２ The frame ceramic green sheet 21G has a planar shape in which a portion corresponding to the cavity CV (FIG. 3) of the ceramic frame 21 is removed. Specifically, the frame ceramic green sheet 21G is formed as a simple sheet by a doctor blade method, and then the portion corresponding to the cavity CV is removed.

[0067] In step ST20 (FIG. 4), the green sheets LG1 to LG9 and the frame ceramic green sheet 21G are stacked to form a laminate SG (FIGS. 5 and 6). The laminate SG includes a laminate 11G formed by stacking the green sheets LG1 to LG9 one on top of another. The laminate 11G will be fired to form the heat sink 11 (FIG. 3).

[0068] Next, at the position where the breaking will be performed as described below, trenches (not shown) may be formed on the surfaces of each of the laminate 11G and the frame ceramic green sheet 21G using mechanical processing with a cutting edge CT (Figure 5) or laser processing with a laser processing device (not shown).

[0069] In step ST30 (FIG. 4), the laminate SG (FIG. 5) is fired. This fires the green sheets LG1 to LG9 and the frame ceramic green sheet 21G. This fires the mixed powder. This fire transforms the laminate SG into a fired body SF (FIG. 7). The firing temperature is, for example, 1100°C or higher and 1400°C or lower. By setting the firing temperature to 1100°C or higher, the laminate SG can be heated to a temperature above the melting point of Cu. In other words, the mixed powder can be heated to a temperature above the melting point of Cu. This allows the formation of a high-quality heat sink 11 containing Cu. On the other hand, by setting the firing temperature to 1400°C or lower, process difficulties caused by an excessively high firing temperature can be avoided.

[0070] Next, a breaking step is performed starting from the trenches, as shown by the broken lines BR (FIG. 7). As a result, the sintered body SF is divided into a plurality of parts. This results in a plurality of sintered bodies SF corresponding to a plurality of packages 51 (FIG. 3) (FIG. 8).

[0071] Next, a lead frame 30 (FIG. 3) is attached to the sintered body SF, thereby obtaining a package 51 (FIG. 3).

[0072] In the above manufacturing method, plating may be performed at an appropriate timing after the firing step. As mentioned above, the above manufacturing method is merely an example, and various modifications may be applied. For example, instead of performing the breaking step on the fired body SF, cutting may be performed on the pre-fired stack SG. Furthermore, although the mounting of the semiconductor element 8 ( FIG. 2 ) is performed after the breaking step according to the above manufacturing method, it may instead be performed after the firing step and before the breaking step.

[0073] Furthermore, in the above manufacturing method, the heat sink 11 is formed from a laminate 11G of multiple green sheets. By increasing the number of stacked green sheets, a thicker heat sink 11 can be easily formed. Alternatively, as a variant, the heat sink 11 may be formed from a single green sheet, simplifying the manufacturing method. As another variant, a manufacturing method that does not include a step of forming a green sheet may be used. For example, instead of forming a green sheet, a molded body may be formed by press-molding powder. By firing the resulting molded body, a heat sink component can be obtained without forming a green sheet. To facilitate press-molding, additives may be added to the mixed powder to be press-molded. The additives are typically materials that substantially disappear by firing at the latest, such as a solvent, a dispersant, a plasticizer, and a resin, or any combination thereof.

[0074] (Effects) According to this embodiment, the heat sink 11 easily achieves good thermal conductivity by containing Cu, and the thermal expansion coefficient can be adjusted by also containing at least one of W and Mo. Furthermore, the Young's modulus can be suppressed by containing silicon oxide, and by not including an excessive amount of silicon oxide, significant adverse effects on the thermal conductivity can be avoided. Among various materials, such as silicon oxide, alumina, zirconia, and titania, which are typical ceramic materials for obtaining ceramic structures, silicon oxide is a preferred material for the particles dispersed in the sintered material portion of the heat sink 11, due to its low Young's modulus and low thermal expansion coefficient, in order to achieve the above-mentioned effects. Although the silicon oxide after firing may be either crystalline or amorphous, amorphous silicon oxide is more preferable due to its lower Young's modulus and thermal expansion coefficient. From the above, the heat sink 11 has a thermal expansion coefficient close to that of ceramic materials and good thermal conductivity, while suppressing a decrease in the reliability of the joint or the ceramic frame 21 (different material member) due to the difference in thermal expansion between the heat sink 11 and the ceramic frame 21 (different material member) to which it is joined.

[0075] As described above, the particle size distribution of the silicon oxide particles preferably has a ratio of 70% or more of the particle size range of 0.2 μm or more and less than 1.0 μm. According to the inventors' studies, when the silicon oxide particles have a particle size range of 0.2 μm or more and less than 1.0 μm, the effect of reducing the Young's modulus of the heat sink 11 is particularly large. It is believed that silicon oxide particles with an excessively large particle size have a small effect of reducing the Young's modulus. In fact, there is a concern that silicon oxide particles with an excessively large particle size may become the starting point for damage to the heat sink 11.

[0076] The heat sink 11 ( FIG. 3 ) may have a side surface P4b that is flush with the outer surface P4a of the ceramic frame 21. A portion of the side surface P4b may be a fracture surface formed during the breaking process ( FIG. 7 ). In this case, it is easy to ensure a mounting area (the area in which a semiconductor element 8 or the like can be mounted) while preventing the area between the outer surface P4a and the side surface P4b from becoming a fracture initiation point for the package 51. Typical examples of the side surface P4b not flush with the outer surface P4a include a configuration in which the outer surface P4a protrudes outward from the side surface P4b, or a configuration in which the outer surface P4a is located inside the side surface P4b. In the former configuration, the protruding portion may become a fracture initiation point. In the latter configuration, since the outer edge of the ceramic frame 21 is inward, the inner edge of the ceramic frame 21 also moves inward as long as the width of the ceramic frame 21 needs to be maintained at a predetermined dimension, resulting in a smaller mounting area.

[0077] Second Embodiment FIG. 9 is a schematic cross-sectional view showing the configuration of a semiconductor module 92 according to the second embodiment. The semiconductor module 92 has a configuration in which the package 51 of the semiconductor module 91 (FIG. 2: first embodiment) is replaced with a package 52 according to the second embodiment. The package 52 has a ceramic frame 29 and a brazing material layer 26 (bonding layer) instead of the ceramic frame 21 (FIG. 2: first embodiment). The ceramic frame 29 is made of ceramic, typically alumina. The brazing material layer 26 is, for example, a silver brazing material. A metallized layer (not shown) is preferably provided on the surface of the ceramic frame 29 facing the brazing material layer 26. Note that a resin adhesive layer may be used as the bonding layer instead of the brazing material layer 26.

[0078] Third Embodiment A heat sink 11 (FIG. 3) without a ceramic frame 21 attached thereto may be produced. For example, in the method for producing a fired body SF (FIG. 7) described in the first embodiment, by omitting the formation of the frame ceramic green sheet 21G (FIG. 5), a heat sink 11 without other components such as the ceramic frame 21 attached thereto can be obtained. Other components may be attached to the heat sink 11 thus obtained, and a package 52 (FIG. 9: second embodiment) can be obtained by attaching a ceramic frame 29 or the like. As a modified example, instead of forming a green sheet, a molded body may be formed by press molding, as also described in the first embodiment.

[0079] <Fourth Preferred Embodiment> Fig. 17 is a schematic cross-sectional view showing the configuration of a semiconductor module 94 according to a fourth preferred embodiment. Fig. 18 is a schematic cross-sectional view showing the configuration of a heat dissipation substrate 54 as a component of the semiconductor module 94 of Fig. 17.

[0080] The semiconductor module 94 has a heat dissipation substrate 54 and a semiconductor element 8 mounted thereon. The heat dissipation substrate 54 has a heat sink 11 and a ceramic insulating layer 24 disposed on the main surface P2 of the heat sink 11. The heat dissipation substrate 54 also has a conductor layer 34 disposed on the ceramic insulating layer 24. The conductor layer 34 is electrically insulated from the heat sink 11 by the ceramic insulating layer 24. The semiconductor element 8 is mounted on the conductor layer 34. A bonding material 291 may be used for this mounting. Note that a bonding wire or the like may be bonded to the conductor layer 34.

[0081] In the fourth embodiment, the thickness of the heat sink 11 is preferably 0.3 mm or more and 3.0 mm or less, more preferably 0.5 mm or more and 1.5 mm or less. If the thickness is too small, the mechanical strength of the heat sink 11 will be insufficient. If the thickness is too large, the thermal resistance will be excessively high. The ceramic insulating layer 24 has a thickness smaller than that of the heat sink 11. The thickness of the ceramic insulating layer 24 is preferably 5 μm or more and 50 μm or less, more preferably 5 μm or more and 20 μm or less. If the thickness is too small, the variation in the thickness of the ceramic insulating layer 24 is likely to become a problem. Specifically, electrical insulation is likely to be insufficient in areas where the thickness is locally small. If the thickness is too large, the thermal resistance will be excessively high. The conductor layer 34 has a thickness smaller than that of the heat sink 11. The thickness of the conductor layer 34 is preferably 5 μm or more and 200 μm or less, more preferably 5 μm or more and 20 μm or less. If the thickness is too small, the variation in thickness of the conductor layer 34 is likely to become a problem, whereas if the thickness is too large, the thermal resistance becomes excessively high.

[0082] The ceramic insulating layer 24 is made of ceramic. The ceramic contains alumina (Al ２ O ３ ), and may contain trace amounts of silica (SiO ２ The ceramic insulating layer 24 may contain, for example, 50 wt % or more of Al as a main component. ２ O ３ powder and SiO ２ % by weight in terms of MnO. When such a powder mixture is used, the firing temperature is, for example, 1150 to 1300°C.

[0083] The conductor layer 34 may contain Cu and at least one high-melting-point metal selected from the group consisting of W and Mo. Furthermore, when the total volume of the conductor layer 34 is defined as 100 vol %, the conductor layer 34 may contain 30 vol % or less of a ceramic. This ceramic is, for example, alumina. Other ceramics may be contained together with or instead of alumina, such as SiO ２ and / or MnO ２ The conductor layer 34 may contain a ceramic. When the conductor layer 34 contains ceramic, the adhesion between the conductor layer 34 and the ceramic insulating layer 24 is improved. The ceramic may also contain silicon oxide particles with an average particle size of 5 nm to 200 nm. The material of the conductor layer 34 may be the same as the material of the heat sink 11 described above. However, for some reason, these materials may be different from each other.

[0084] FIG. 19 is a schematic partial cross-sectional view illustrating a step in a manufacturing method of a heat dissipation substrate 54. In this step, instead of the laminate SG (FIG. 6: Embodiment 1) including the laminate 11G, a laminate SG4 including the laminate 11G is formed. As in the first embodiment, a single green sheet may be used instead of the laminate 11G. The laminate SG4 further includes a green sheet 24G that becomes the ceramic insulating layer 24 when fired. Instead of laminating the green sheet 24G on the laminate 11G, a ceramic paste layer that becomes the ceramic insulating layer when fired may be formed by a printing method. The laminate SG4 further includes a green sheet 34G that becomes the conductor layer 34 when fired. Instead of the green sheet 34G, a conductor paste layer that becomes the conductor layer 34 when fired may be formed by a printing method. The heat dissipation substrate 54 is obtained by firing the laminate SG4.

[0085] Note that the configuration other than that described above is substantially the same as that of the first embodiment, and therefore the same or corresponding elements are given the same reference numerals and their description will not be repeated.

[0086] Examples and Comparative Examples A single heat dissipation member such as the heat dissipation plate 11 of the third embodiment was fabricated and evaluated. First, the raw material composition, i.e., the composition of the mixed powder, the raw material particle size, and the Cu introduction method are shown in Table 1 below.

[0087]

[0088] Amorphous silicon oxide was used. In the "Cu introduction method" column in Table 1, "impregnation" indicates that Cu was introduced into a W or Mo porous body by impregnation. "Powder mixing" indicates that Cu powder was mixed in the powder mixing step to prepare the powder to be fired. The firing temperature was 1250°C.

[0089] In Table 1, the composition of the heat dissipation member is expressed based on the total weight of Cu, W, and Mo. However, Cu, W, Mo, and silicon oxide (SiO ２ ) and converted to the total weight as shown in Table 2 below.

[0090]

[0091] The compositions and evaluation results of the heat dissipating members obtained using the above raw materials are shown in Table 3 below.

[0092]

[0093] According to preliminary studies by the inventors, the compositional variation from the raw material composition (composition of the mixed powder) of the heat dissipation member to the composition after firing (composition of the heat dissipation member) was sufficiently small for Cu, W, and Mo. Therefore, the raw material composition values ​​listed in Table 1 were adopted for these elements. On the other hand, the silicon oxide content estimated from the aforementioned 2000x electron microscope photograph of the cross section of the heat dissipation member was significantly reduced from the silicon oxide content in the mixed powder. As a precaution, a backscattered electron image at 25000x magnification was also confirmed, but no silicon oxide particles other than those recognized in the above electron microscope photograph were observed. Furthermore, in qualitative analysis using an electron probe microanalyzer (EPMA) in areas where silicon oxide particles were not observed, significant amounts of Si element were not detected. Therefore, it is thought that silicon oxide may have been released from the heat dissipation member during firing. Therefore, the silicon oxide content after firing was calculated using image data of the electron microscope photograph, as described above. The content is SiO ２ It is expressed in terms of conversion.

[0094] Referring to Table 3, in Comparative Example 5, the sintering state was insufficient, that is, the sintered body was not dense and many pores were present in the sintered body. ２ It is thought that the particle size of the powder raw material is related. The detailed reason is unknown, but SiO ２ This phenomenon becomes more pronounced when the particle size of the powder is greater than 200 nm. Although not shown in Table 3, the same results were confirmed when Mo in Comparative Example 5 was replaced with W.

[0095] In addition, in Comparative Examples 1 to 4, unlike the present embodiment, the impregnation method was used to introduce Cu elements.

[0096] Furthermore, "silicon oxide particle size distribution" refers to the particle size distribution based on the number of particles, based on electron microscope observation of silicon oxide particles. Specifically, "0.2-1.0" indicates the proportion of particle sizes in the particle size range of 0.2 μm or more and less than 1.0 μm, and "1.0-2.0" indicates the proportion of particle sizes in the particle size range of 1.0 μm or more and less than 2.0 μm. From these results, in the particle size distribution based on the number of particles in the particle size range of 0.2 μm or more and less than 2.0 μm, the proportion of particle sizes in the particle size range of 0.2 μm or more and less than 1.0 μm was 70% or more in each example. To provide a supplementary explanation of the investigations conducted by the present inventors, in the particle size distribution based on the number of particles in the particle size range of 0.2 μm or more and less than 3.0 μm, the proportion of particle sizes in the particle size range of 0.2 μm or more and less than 1.0 μm was 70% or more in each example. To further explain, in the particle size distribution based on the number of particles in the particle size range of 0.2 μm or more and less than 5.0 μm, the proportion of the particle size range of 0.2 μm or more and less than 1.0 μm was 70% or more in each Example. To further explain, in the particle size distribution based on the number of particles in the particle size range of 0.2 μm or more and less than 10 μm, the proportion of the particle size range of 0.2 μm or more and less than 1.0 μm was 70% or more in each Example.

[0097] Furthermore, in all examples, each of the silicon oxide particles in the heat dissipation member had a particle size of less than 10 μm. It is believed that the number of silicon oxide particles with a particle size of 10 μm or more is preferably small, and more preferably substantially zero. This further reduces the risk of silicon oxide particles becoming a starting point for damage to the heat dissipation plate 11. Further, to provide additional information on the investigations conducted by the present inventors, each of the silicon oxide particles in the heat dissipation member had a particle size of less than 5.0 μm. It is believed that the number of silicon oxide particles with a particle size of 5.0 μm or more is preferably small, and more preferably substantially zero. This further reduces the risk of silicon oxide particles becoming a starting point for damage to the heat dissipation plate 11. To further supplement the analysis conducted by the present inventors, each of the silicon oxide particles in the heat dissipation member had a particle size of less than 3.0 μm. It is believed that the number of silicon oxide particles with a particle size of 3.0 μm or more is preferably small, and more preferably substantially zero. This further reduces the risk of silicon oxide particles becoming a starting point for damage to the heat dissipation plate 11. To further explain, in all examples, the silicon oxide particles in the heat dissipation member each had a particle size of less than 2.0 μm. It is considered preferable that the number of silicon oxide particles with a particle size of 2.0 μm or more is small, and more preferably, substantially zero. This further reduces the risk of silicon oxide particles becoming the starting point for damage to the heat dissipation plate 11.

[0098] In Table 3 above, the "Sintered state" column shows the results of an evaluation of whether or not a sintered state sufficient to withstand use as a heat dissipation component was obtained. The "thermal expansion coefficient" was calculated based on thermal expansion between room temperature and 100°C. In addition, as an index of Young's modulus, the stress value when a bending strain of 1% was applied was measured. The smaller the stress value, the smaller the Young's modulus. The bending strain value was measured using the JIS R1602 three-point bending strain gauge method.

[0099] Figures 10, 11, and 12 are electron microscope photographs of the cross sections of heat dissipation members of Examples 2, 3, and 9, respectively. In these photographs, the white areas represent tungsten, the gray areas represent copper, and the black areas represent silicon oxide particles. The silicon oxides of Examples 1 to 14 were all amorphous. Figures 13(a) and 13(b) are graphs showing the relationship between Cu content and stress at 1% strain. For the heat dissipation plate containing Cu and W in Figure 13(a) and the heat dissipation plate containing Cu and Mo in Figure 13(b), the stress at 1% strain was lower in the Examples (black circle markers) than in the Comparative Examples (triangle markers). Figures 14(a) and 14(b) are graphs showing the relationship between thermal expansion coefficient and stress at 1% strain. For example, when comparing the same thermal expansion coefficient data, the Examples (black circle markers) had smaller stress values ​​than the Comparative Examples (triangle markers).

[0100] FIG. 15 is a schematic diagram showing the microstructure of a heat dissipation member containing Cu and W but without dispersed silicon oxide particles. In this structure, Cu penetrates the voids between sintered W particles. In this case, direct bonding between W particles with a high Young's modulus is easily formed. As a result, the Young's modulus of the heat dissipation member is also high. In contrast, FIG. 16 is a schematic diagram showing the microstructure of a heat dissipation member containing Cu and W and with dispersed silicon oxide particles. In this structure, the fine silicon oxide particles located between W particles inhibit direct bonding between W particles with a high Young's modulus, and Cu, with a low Young's modulus, easily penetrates between the W particles. As a result, the Young's modulus of the heat dissipation member is reduced. While fully utilizing the effects of the silicon oxide particles, the fine silicon oxide particles allow the silicon oxide content to be reduced to a level that does not significantly adversely affect the thermal conductivity of the heat dissipation member. The same applies when Mo particles are used instead of W particles in FIGS. 15 and 16 .

[0101] In Table 3, the composition of the heat dissipation member is expressed based on the total weight of Cu, W, and Mo. However, Cu, W, Mo, and silicon oxide (SiO ２) and the results are as shown in Table 4 below.

[0102]

[0103] 11: Heat sink (heat dissipation member) 11G, SG: Laminated body 21, 29: Ceramic frame 21G: Frame ceramic green sheet 24: Ceramic insulating layer 26: Brazing material layer 34: Conductive layer 51, 52: Package 54: Heat dissipation substrate LG1 to LG9: Green sheets SF: Fired body

Claims

1. a sintered material portion containing copper and at least one of tungsten and molybdenum; a plurality of silicon oxide particles dispersed in the sintered material portion; A heat dissipation member comprising: In the heat dissipation member, the copper content is M with respect to the total weight of copper, tungsten, and molybdenum. Ｃｕ weight percent, and the tungsten content is M Ｗ weight percent, and the molybdenum content is M Ｍｏ percent by weight, and the silicon oxide content is SiO ２ Converted to M ＳｉＯ２ is weight percent, 0.9≧M Ｃｕ / (M Ｃｕ +M Ｗ +M Ｍｏ ) ≧0.045, and 0.01≧M ＳｉＯ２ / (M Ｃｕ +M Ｗ +M Ｍｏ )≧0.0003 A heat dissipation member that is filled with

2. The heat dissipation member according to claim 1, A heat dissipation member, wherein in a particle size distribution based on the number of particles within the particle size range of 0.2 μm or more and less than 10 μm, the proportion of particles within the particle size range of 0.2 μm or more and less than 1.0 μm is 70% or more.

3. The heat dissipation member according to claim 2, A heat dissipation member, wherein each of the plurality of silicon oxide particles has a particle size of less than 10 μm.

4. The heat dissipation member according to claim 1, A heat dissipation member, wherein in a particle size distribution based on the number of particles within the particle size range of 0.2 μm or more and less than 5 μm, the proportion of particles within the particle size range of 0.2 μm or more and less than 1.0 μm is 70% or more.

5. The heat dissipation member according to claim 4, A heat dissipation member, wherein each of the plurality of silicon oxide particles has a particle size of less than 5 μm.

6. The heat dissipation member according to claim 1, A heat dissipation member, wherein in a particle size distribution based on the number of particles within the particle size range of 0.2 μm or more and less than 3 μm, the proportion of particles within the particle size range of 0.2 μm or more and less than 1.0 μm is 70% or more.

7. The heat dissipation member according to claim 6, A heat dissipation member, wherein each of the plurality of silicon oxide particles has a particle size of less than 3 μm.

8. The heat dissipation member according to claim 1, A heat dissipation member, wherein in a particle size distribution based on the number of particles within the particle size range of 0.2 μm or more and less than 2 μm, the proportion of particles within the particle size range of 0.2 μm or more and less than 1.0 μm is 70% or more.

9. The heat dissipation member according to claim 8, A heat dissipation member, wherein each of the plurality of silicon oxide particles has a particle size of less than 2 μm.

10. The heat dissipation member according to any one of claims 1 to 9, 0.80≧M Ｃｕ / (M Ｃｕ +M Ｗ +M Ｍｏ )≧0.15 A heat dissipation member that is filled with

11. The heat dissipation member according to any one of claims 1 to 9, A heat dissipation member, wherein the balance other than copper, tungsten, molybdenum, and silicon oxide in the heat dissipation member is less than 0.5 weight percent of the total weight.

12. The heat dissipation member according to any one of claims 1 to 9, M Ｍｏ = 0, and 0.806≧M Ｃｕ / (M Ｃｕ +M Ｗ )≧0.075 A heat dissipation member that is filled with

13. The heat dissipation member according to any one of claims 1 to 9, M Ｗ = 0, and 0.887≧M Ｃｕ / (M Ｃｕ +M Ｍｏ )≧0.133 A heat dissipation member that is filled with

14. A method for manufacturing a heat dissipation member according to any one of claims 1 to 9, comprising: At least one powder of tungsten having an average particle size of 0.5 μm or more and 10 μm or less and molybdenum having an average particle size of 0.5 μm or more and 10 μm or less, copper powder having an average particle size of 1.5 μm or more and 5.0 μm or less, and SiO powder having an average particle size of 7 nm or more and 200 nm or less. ２ and a powder containing copper and tungsten, and molybdenum, to form a mixed powder, wherein the mixed powder contains copper in an amount of M based on the total weight of copper, tungsten, and molybdenum. Ｃｕ（Ｐ） weight percent, and the tungsten content is M Ｗ（Ｐ） weight percent, and the molybdenum content is M Ｍｏ（Ｐ） percent by weight, and the silicon oxide content is SiO ２ Converted to M ＳｉＯ２（Ｐ） is weight percent, 0.9≧M Ｃｕ（Ｐ） / (M Ｃｕ（Ｐ） +M Ｗ（Ｐ） +M Ｍｏ（Ｐ） ) ≧0.045, and 0.03≧M ＳｉＯ２（Ｐ） / (M Ｃｕ（Ｐ） +M Ｗ（Ｐ） +M Ｍｏ（Ｐ） )≧0.001 is satisfied, and the manufacturing method further comprises: A method for manufacturing a heat dissipation member, comprising the step of heating the mixed powder to a temperature equal to or higher than the melting point of copper.

15. The method for manufacturing a heat dissipation member according to claim 14, forming at least one green sheet containing the mixed powder and a resin; The method for manufacturing a heat dissipation member, wherein the step of heating the mixed powder is carried out by firing the at least one green sheet.

16. The method for manufacturing a heat dissipation member according to claim 15, the at least one green sheet is a plurality of green sheets; further comprising a step of forming a laminate by stacking the plurality of green sheets on one another; The method for manufacturing a heat dissipation member, wherein the step of heating the mixed powder is carried out by firing the laminate.

17. The heat dissipation member according to any one of claims 1 to 9; A ceramic frame; A package comprising: the heat dissipation member has a heat dissipation surface and a main surface opposite to the heat dissipation surface, The ceramic frame is disposed on the main surface of the heat dissipation member, and has an inner surface surrounding a cavity and an outer surface opposite the inner surface.

18. The heat dissipation member according to any one of claims 1 to 9; a ceramic insulating layer; A substrate comprising: the heat dissipation member has a heat dissipation surface and a main surface opposite to the heat dissipation surface, The ceramic insulating layer is disposed on the main surface of the heat dissipation member.