JP5513058B2 - Composite member manufacturing method, composite member, heat dissipation member, and semiconductor device - Google Patents

Description

  The present invention relates to a composite member in which magnesium (so-called pure magnesium) or a magnesium alloy and a non-metallic inorganic material such as SiC are combined, a heat radiating member composed of the composite member, a semiconductor device including the heat radiating member, and a composite member It is related with the manufacturing method. In particular, the present invention relates to a method of manufacturing a composite member that can manufacture a high-quality composite member even if it is large.

  As a constituent material of a heat dissipation member (heat spreader) of a semiconductor element, a composite material of a metal and a non-metallic inorganic material (typically ceramics) such as Al-SiC is used in addition to a material made of only a metal material such as copper . In recent years, a composite material using magnesium (Mg) or its alloy as a base material, which is lighter than aluminum (Al), has been studied mainly for the purpose of reducing the weight of the heat dissipation member (see Patent Document 1).

JP 2006-299304 A

  In recent years, as various types of electronic devices such as personal computers and portable electronic devices become more functional and have higher density packaging, the number of semiconductor elements and peripheral components mounted on one wiring board tend to increase. Correspondingly, a large heat radiating member is desired. Although Patent Document 1 discloses a plate material having a size of 50 mm × 100 mm as a composite material of Mg or Mg alloy and SiC or the like, further enlargement is desired. However, conventionally, a method for producing a larger composite material has not been sufficiently studied. As a result of investigations by the present inventors, it has been found that, as will be described later, in a large-sized composite material, a portion where defects are concentrated is formed and the quality may be lowered.

  Accordingly, one of the objects of the present invention is a composite member of Mg or Mg alloy and a non-metallic inorganic material, and a method of manufacturing a composite member capable of manufacturing a high-quality composite member even if it is large Is to provide. Another object of the present invention is to provide a composite member of Mg or Mg alloy and a non-metallic inorganic material, which is large but has a high quality. Furthermore, the other object of this invention is to provide the heat radiating member which consists of the said composite member, and a semiconductor device provided with this heat radiating member.

  In order to examine the problems of the prior art, the present inventors have a size exceeding 50 mm × 100 mm, specifically, a plate having a size capable of taking a circular region having a diameter of more than 50 mm. Infiltration method as described in Patent Document 1 (Mg Mg or Mg alloy (hereinafter referred to as molten Mg) melted in a non-metallic inorganic material filled in a mold) An infiltration plate in which Mg or an Mg alloy and a non-metallic inorganic material were combined was produced by a method of solidifying the material and slowly cooled in an atmosphere furnace. Then, when the obtained plate-shaped composite member was cut in the thickness direction and the cross section was observed, defects were concentrated in the central portion of the composite member, and defects were found in the peripheral portion. Virtually not seen. More specifically, large voids (pores) due to shrinkage nests (inner shrinkage nests) were observed inside the central portion. Further, outer shrinkage nests and irregularities were also found on the surface of the central portion of the composite member. The reason why the defect is locally generated in this way is estimated as follows.

  When the infiltration plate is gradually cooled, solidification of molten Mg proceeds so as to converge from the peripheral edge toward the central portion and from the surface toward the inside. Then, a hot spot (an unsolidified metal region surrounded by a solidified metal) may be generated inside the central portion. Here, the volume of Mg or Mg alloy decreases by about 4% when solidified. For this reason, solidification proceeds so that unsolidified Mg or Mg alloy compensates for the volume reduction due to solidification. Then, when the undissolved metal in the hot spot in the central portion is solidified, Mg or Mg alloy is insufficient due to the above-described compensation, and the above-described compensation is not performed. As a result, it is considered that defects are hardly generated in the peripheral portion of the plate-shaped composite member, and a large gap is likely to be generated in the central portion. Further, even if the molten Mg in the surface portion has already solidified in the central portion, the inner unsolidified portion is pulled inside due to the volume reduction when solidified. For this reason, it is considered that an outer shrinkage nest is generated on the surface of the central portion, and unevenness along the outer shape of the nonmetallic inorganic material used as a raw material is generated by the surface shrinkage. And when the above big defects concentrate in the inside of a composite member, mechanical characteristics, such as intensity | strength, and thermal characteristics, such as thermal conductivity, will fall. Here, when a plate-shaped composite member is used as a heat radiating member of a semiconductor element, a semiconductor element or the like is usually mounted on the central portion of the composite member. Therefore, a composite member in which defects are concentrated on a portion where a semiconductor element is mounted is not suitable for a heat dissipation member. In addition, if surface defects such as the above-described shrinkage nests exist outside the composite member, surface properties are deteriorated and dimensional accuracy is reduced. In particular, composite materials of non-metallic inorganic materials such as SiC and metals contain SiC that is easier to cool than molten metal, so the above internal defects compared with the case of forming a metal-only casting using only molten metal And surface defects are likely to occur.

  In order to reduce the above defects, the present inventors do not converge the cooling of the infiltrating plate from the peripheral edge toward the central portion, but instead from one edge of the infiltrating plate to the other opposite edge. The knowledge that it is effective to cool in one direction so that it goes is obtained. The present invention is based on the above findings.

The method for producing a composite member according to the present invention is a method for producing a composite member in which magnesium or a magnesium alloy and a nonmetallic inorganic material are composited, and includes the following composite step and cooling step.
Compounding step: A step of infiltrating molten magnesium or a magnesium alloy into an aggregate of nonmetallic inorganic materials housed in a mold to form an infiltrating plate.
Cooling step: a step of cooling the infiltrating plate in one direction from the opposite side of the molten magnesium or magnesium alloy supply side to solidify the molten magnesium or magnesium alloy.
Typically, in the composite step, the molten magnesium or magnesium alloy is supplied to the aggregate by its own weight from the gravity direction, that is, from the vertical direction upper side to the vertical direction lower side. In the cooling step, the infiltrating plate is cooled in one direction from the lower side in the vertical direction toward the upper side in the vertical direction, and the molten magnesium or magnesium alloy is solidified.

  The above-described method for producing a composite member of the present invention can of course be used for the production of a small composite member capable of taking a circular region having a diameter of 50 mm or less, and in particular, taking a circular region having a diameter of more than 50 mm. This can be suitably used in the case of manufacturing a composite member having a size that can be used. In this case, the infiltration plate may have a size capable of taking a circular region having a diameter of more than 50 mm.

  Moreover, according to the said manufacturing method of this invention composite member, the following this invention composite member is obtained, for example. The composite member of the present invention is a plate material in which magnesium or a magnesium alloy and a nonmetallic inorganic material are combined, and has a size capable of taking a circular region having a diameter of more than 50 mm. Then, a cross section in the thickness direction of the composite member is taken with a straight line passing through the center of gravity when the composite member is viewed in plan as a cutting line. In this cross section, a region in the range of up to 10% of the length of the cutting line along the longitudinal direction of the cutting line with the center of gravity as the center is taken as a central region. When an arbitrary small area of 1 mm × 1 mm is taken from this central area, and the ratio of the area of the defect portion to the area of the small area is an area ratio, the area ratio of the composite member is 10% or less.

  In the production method of the present invention, if the molten Mg supply direction is, for example, the gravitational direction as described above, the undissolved molten Mg in the infiltration plate is always on the upper side in the vertical direction (typically From the mold opening side to the lower side in the vertical direction (typically, the bottom surface side of the mold). In this state, when the infiltration plate is cooled in one direction from the vertical lower side to the vertical upper side, when the lower molten Mg is solidified, the upper unsolidified Mg becomes lower. Are sequentially supplied to the side. That is, solidification proceeds toward the upper side in the vertical direction while the volume decrease when the molten Mg on the lower side in the vertical direction solidifies is sequentially compensated by the unsolidified molten Mg present on the upper side in the vertical direction. Therefore, in this manufacturing method, as described above, there is a large defect in the central portion (the portion including the central region) of the composite member as in the case where the cooling is converged from the periphery of the infiltration plate toward the central portion. There is no concentration. As described above, according to the composite member of the present invention obtained by setting the cooling direction to one specific direction, large defects do not exist intensively and are of high quality. In addition, the composite member of the present invention manufactured by the manufacturing method of the present invention is a large plate material having a size capable of taking a circular region having a diameter of more than 50 mm, but is in the central portion including the central region. Defects such as voids associated with shrinkage nests are small. In this way, the composite member of the present invention in which no large defects exist locally is excellent in mechanical properties such as strength and thermal properties such as thermal conductivity, and has a sufficient mounting area for semiconductor elements and peripheral components. Can do. Therefore, the composite member of the present invention can be suitably used as a heat dissipation member for the semiconductor element.

Hereinafter, the present invention will be described in more detail.
[Composite material]
As a form of the composite member of the present invention, a form of only a substrate made of a composite material in which Mg or an Mg alloy and a non-metallic inorganic material are composited, the substrate, and a metal coating layer covering at least one surface of the substrate are provided. Form. First, the substrate will be described.

<Metal component>
The metal component in the substrate is so-called pure magnesium composed of 99.8% by mass or more of Mg and impurities, or a magnesium alloy composed of additive elements and the balance Mg and impurities. When the metal component is pure magnesium, compared to the case of an alloy, (1) the thermal conductivity of the composite member can be improved, and (2) the crystallized product is deposited unevenly during solidification. Since it does not easily occur, there is an advantage that it is easy to obtain a composite member having a uniform structure. When the metal component is a magnesium alloy, the liquidus temperature is lowered, so that the temperature at the time of melting can be lowered and the corrosion resistance and mechanical properties (strength, etc.) of the composite member can be improved. Examples of the additive element include at least one of Li, Ag, Ni, Ca, Al, Zn, Mn, Si, Cu, and Zr. Since these elements cause a decrease in thermal conductivity when the content increases, the total content is preferably 20% by mass or less (the metal component is 100% by mass; the same applies to the content of the additive element). In particular, Al is preferably 3% by mass or less, Zn is 5% by mass or less, and other elements are each preferably 10% by mass or less. Addition of Li has the effect of reducing the weight of the composite member and improving the workability. Known magnesium alloys such as AZ, AS, AM, ZK, ZC, and LA may be used. A metal raw material is prepared so as to have a desired composition.

<Non-metallic inorganic material>
"composition"
Examples of the non-metallic inorganic material in the substrate include those having a thermal expansion coefficient smaller than Mg, excellent thermal conductivity, and hardly reacting with Mg. As such a non-metallic inorganic material, a ceramic such as SiC is representative. Others, Si ₃ N ₄ , Si, MgO, Mg ₃ N ₂ , Mg ₂ Si, MgB ₂ , MgCl ₂ , Al ₂ O ₃ , AlN, CaO, CaCl ₂ , ZrO ₂ , diamond, graphite, h-BN, c -At least one of BN, B ₄ C, Y ₂ O ₃ , and NaCl is included. In particular, SiC has (1) a coefficient of thermal expansion of about 3 ppm / K to 4 ppm / K, which is close to the coefficient of thermal expansion of semiconductor elements and peripheral components, and (2) has a particularly high thermal conductivity among non-metallic inorganic materials. (Single crystal: about 390W / m · K to 490W / m · K), (3) Powders and sintered bodies of various shapes and sizes are commercially available, (4) Advantages such as high mechanical strength Have You may contain the said multiple types of nonmetallic inorganic material. For example, a part of the non-metallic inorganic material is allowed to exist as a network part described later.

<< Content and form of existence >>
When the content of the nonmetallic inorganic material in the substrate is 20% by volume when the substrate is 100% by volume, the thermal conductivity κ is high and the thermal expansion coefficient (linear thermal expansion coefficient) α is small. It can be a substrate. The greater the content of the non-metallic inorganic material in the substrate, the higher the thermal conductivity κ and the smaller the thermal expansion coefficient α, and the semiconductor element (about 4 ppm / K to 8 ppm / K (e.g., Si: 4.2 ppm / K, GaAs: 6.5 ppm / K) and its peripheral components (insulating substrate: about 4.5 ppm / K, silicon package: about 3 ppm / K, alumina package: 6.5 ppm / K). Accordingly, the content is preferably 50% by volume or more, particularly 70% by volume or more, more preferably 80% by volume or more, and particularly preferably 85% by volume or more. Although no particular upper limit is set, it is considered that about 80% by volume to 90% by volume is practical considering the infiltration property of molten Mg, industrial productivity, and the like.

  The content of the nonmetallic inorganic material in the substrate is substantially equal to the amount of the raw material. Further, for example, when a powder of a nonmetallic inorganic material is used as a raw material, the substrate has a dispersed form in which particles of the nonmetallic inorganic material are dispersed in a discrete manner in the metal component of Mg or Mg alloy. Alternatively, when a molded body of the non-metallic inorganic material is used as a raw material, in particular, a porous molded body having a network part that bonds particles by heating the powder molded body, the substrate has the network part. In addition, the gap (pores) of the molded body of the nonmetallic inorganic material is impregnated with Mg or Mg alloy, or the network portion of the molded body is collapsed during infiltration to become the dispersed form. To do. By appropriately adjusting the sintering conditions and the like, the presence state of the nonmetallic inorganic material in the substrate can be changed. The presence state of the nonmetallic inorganic material in the substrate, for example, the presence or absence and shape of the network portion can be confirmed, for example, by observing the cross section of the substrate with an optical microscope or a scanning electron microscope (SEM). The amount and shape of the raw nonmetallic inorganic material may be appropriately selected so that the substrate has desired thermal characteristics.

<Size and shape>
One of the features of the composite member (substrate) of the present invention is that it is relatively large. One form includes a rectangular shape having a short side of more than 50 mm and a long side of more than 100 mm. Since the composite member (substrate) is thus large, for example, when the composite member is used as a constituent material of a heat dissipation member of a semiconductor device, many semiconductor elements and peripheral components can be mounted. The size and shape of the substrate can be appropriately selected and are not particularly limited. For example, in addition to the rectangular shape, a square shape or a circular shape may be used. A mold is prepared so that the substrate has a desired size and shape. Further, it is considered that the thickness of only the substrate excluding the metal coating layer is preferably 10 mm or less, particularly 5 mm or less, as a constituent material of the heat dissipation member of the semiconductor element.

<Internal organization>
One of the features of the composite member (substrate) of the present invention is that it is large as described above, but defects are not concentrated in the central region, specifically, the above selected from the central region. The area ratio of the defective portion in the small region is 10% or less. The said defect part is a component except Mg or an Mg alloy and a nonmetallic inorganic material in the said small area | region, and a space | gap is mentioned typically.

  The composite member (substrate) of the present invention can satisfy the above area ratio of 5% or less, particularly 2% or less by appropriately adjusting the cooling conditions as described later. Since it is desirable that no defects exist, there is no particular lower limit for the area ratio. When cooling in one specific direction is not performed as in the production method of the present invention, a composite member (substrate) having an area ratio of more than 10% is obtained. That is, this substrate has many defects in the central region, and this defect is very large. More specifically, this substrate has defects of a size that can be visually confirmed, and such large defects are unevenly distributed in the central region. On the other hand, the composite member (substrate) of the present invention has few defects in the central region, and this defect has a size that is difficult to visually confirm (the maximum length of the defect is about 0.1 mm (100 μm) or less). It is. The composite member (substrate) of the present invention does not have a large defect locally, and is very fine even if a defect exists. Therefore, the composite member (substrate) of the present invention is superior in mechanical properties such as strength and thermal properties such as thermal conductivity, compared to the case where large defects are concentrated in part. In the composite member of the present invention, defects are not intensively present not only in the central region but also over substantially the entire cross section. That is, the composite member of the present invention has a uniform structure with no defects unevenly distributed throughout. In the case where the composite member has a metal coating layer on the surface of the substrate made of the composite material, the central region is extracted only for the substrate excluding the metal coating layer.

<Surface condition>
The composite member (substrate) of the present invention is free from internal defects as described above, and has few surface defects such as outer shrinkage nests. Therefore, the composite member (substrate) of the present invention has less surface irregularities and can satisfy, for example, a surface roughness Ra of 2.5 μm or less. Also, because the surface roughness of the substrate itself is small, even when a metal coating layer is formed on the substrate simultaneously with the composite by molten Mg, especially when a thin metal coating layer having a thickness of 1 mm or less is formed. The surface of the metal coating layer is also smooth, and the surface roughness Ra of the metal coating layer can satisfy 2.5 μm or less. The surface of the metal coating layer can also be smoothed by forming the metal coating layer thick and performing machining such as polishing. However, since the metal coating layer may be thin as will be described later, performing the above polishing causes a decrease in yield and an increase in polishing cost.

  In addition, the composite member (substrate) of the present invention has less surface defects as described above, so that deformation is small and dimensional accuracy is excellent. For example, the composite member (substrate) of the present invention can satisfy a difference between the maximum thickness and the minimum thickness of the composite member of 0.2 mm or less, particularly 0.05 mm (50 μm) or less in the cross section of the central region described above. Even when a metal coating layer is provided on the surface of the substrate, since the surface defects are small as described above, the difference can satisfy 0.2 mm or less, particularly 0.05 mm (50 μm) or less.

<Thermal characteristics>
Although it depends on the content of the nonmetallic inorganic material, the presence form of the nonmetallic inorganic material, the composition of the metal component, etc., the composite member (substrate) of the present invention is also excellent in thermal characteristics. Specifically, for example, high thermal conductivity such that the thermal conductivity κ satisfies 180 W / m · K or more, particularly 200 W / m · K or more, 250 W / m · K or more, especially 300 W / m · K or more. Can have. The coefficient of thermal expansion is 3.5ppm / K (3.5 × ^10-6 / K) or more and 20ppm / K (20 × ^10-6 / K) or less, especially 4.0ppm / K (4.0 × ^10-6 / K) or more. A substrate having a relatively small thermal expansion coefficient satisfying 12 ppm / K (12 × 10 ⁻⁶ / K) or less can be obtained. As the content of non-metallic inorganic material having excellent thermal conductivity such as SiC increases, the composite material (substrate) tends to have a higher thermal conductivity κ and a smaller thermal expansion coefficient α.

<Metal coating layer>
<Composition, organization>
The heat radiating member and the semiconductor element or a cooling device for cooling the semiconductor element may be joined by solder. The substrate made of the above composite material does not have good wettability with solder, and it is necessary to improve the wettability with solder by plating with Ni or the like. The plating is preferably electroplating in consideration of productivity, but it is difficult to perform electroplating because many nonmetallic inorganic materials have high electrical insulation. Therefore, by providing a metal coating layer on at least one surface of the substrate and using the metal coating layer as a base for the electroplating, the substrate can be easily electroplated with Ni or the like. In addition, since the surface of the substrate is smooth as described above, the surface of the metal coating layer can be smooth even when the metal coating layer is thin. As a result, the plating can be formed to a uniform thickness.

  The constituent metal of the metal coating layer may be a metal having an electrical conductivity sufficient to provide electrical conduction necessary for electroplating, and may have a composition different from or the same as the metal component of the substrate made of the composite material. In particular, in the case of the same composition, when a metal coating layer is formed with molten Mg and combined, a composite member having the metal coating layer can be produced with high productivity. In this case, in the obtained composite member, the metal component in the substrate and the metal constituting the metal coating layer have a continuous structure (cast structure).

  When the metal component of the substrate and the constituent metal of the metal coating layer have different compositions, the constituent metal of the metal coating layer is, for example, an Mg alloy having a composition different from the metal component of the substrate, or other than Mg and Mg alloy Selected from the group consisting of metals, for example, Al, Cu, Ni with a purity of 99% or more and alloys containing Al, Cu, Ni as the main component (alloys containing more than 50% by mass of Al, Cu, Ni) One kind of metal is mentioned.

<Formation point>
The said metal coating layer should just exist in the surface where plating is required among the surfaces which comprise the said board | substrate. Specifically, a metal coating layer is provided on at least one of the mounting surface on which the semiconductor element is mounted and the cooling surface that faces the mounting surface and contacts the cooling device. A metal coating layer may be provided on the entire surface including the end surface of the substrate (the surface connecting the mounting surface and the cooling surface). In addition to being able to perform electroplating, the composite member provided with the metal coating layer can enhance corrosion resistance and can increase the commercial value of the composite member because of its smooth surface and excellent appearance.

"thickness"
When the thickness of each metal coating layer is too thick, the coefficient of thermal expansion is particularly large, so 2.5 mm or less, particularly 1 mm or less, more preferably 0.5 mm or less, 1 μm or more, particularly 50 μm or more and 500 μm (0.5 mm) or less. If present, it is considered that the metal coating layer is sufficiently damaged while the composite member is sufficiently transported and mounted. In view of productivity as described above, it is preferable to form a thin metal coating layer without performing polishing or the like. Moreover, even if it is thin, the metal coating layer which is excellent in surface property can be formed by utilizing this invention manufacturing method which performs cooling of one specific direction as mentioned above.

  In particular, when the content of the nonmetallic composite material in the substrate is more than 70% by volume, or when it is 50% by volume or more and has a network portion, the substrate has a thickness of about 10 mm or less. The thermal expansion coefficient of a single substance can be set to about 4 ppm / K to 6 ppm / K. For a substrate having such a small coefficient of thermal expansion, even if each of the metal coating layers is formed to a thickness of about 1 mm (total thickness of about 2 mm), the entire composite member including the substrate and the metal coating layer is formed. The thermal expansion coefficient can be 8 ppm / K or less.

<Application>
The said composite member can be utilized suitably for a heat radiating member. This heat radiating member can be suitably used for a heat radiating member of a semiconductor element. In particular, this heat radiating member has no local defects, has few large internal defects, and has few surface defects and excellent surface properties. Therefore, the heat radiating member is expected to have high quality and high commercial value. Moreover, since this heat radiating member is large in size as well as having excellent mechanical strength such as strength and thermal characteristics due to the small number of defects as described above, many semiconductor elements and their peripheral components can be mounted. A semiconductor device including such a heat radiating member and a semiconductor element mounted on the heat radiating member can be suitably used for parts of various electronic devices.

[Production method]
The composite member basically includes a nonmetallic inorganic material aggregate disposed in a mold, and molten Mg or an Mg alloy (hereinafter referred to as molten Mg) is infiltrated into the aggregate to form an infiltration plate. It is formed by cooling this infiltrated plate with a specific one-way cooling and solidifying Mg or Mg alloy.

<Mold>
In producing a plate-shaped composite member such as a rectangular plate, a mold used in the composite process typically includes a bottom surface portion and a side wall portion standing from the bottom surface portion, and the side facing the bottom surface portion is open. A box-shaped one is available. When producing a rectangular plate-shaped composite member, a rectangular parallelepiped box can be suitably used as a mold. After the aggregate of nonmetallic inorganic materials is stored in the box-shaped mold, typically, the mold is arranged so that the bottom surface side of the mold is the lower side in the vertical direction and the opening side of the mold is the upper side in the vertical direction. Then, for example, using the dead weight of molten Mg, the molten Mg flows from the opening side of the mold (the molten Mg supply side) toward the bottom surface side (the side opposite to the supply side), and the aggregate Infiltrate into. By doing so, the infiltrating plate can be easily manufactured, for example, even at atmospheric pressure without pressurizing molten Mg.

  Alternatively, molten Mg can be infiltrated from the bottom surface side of the mold toward the upper opening side. That is, the bottom side of the mold can be the molten Mg supply side, and the mold opening side can be the opposite side of the supply side. For example, a molten Mg pouring port is provided on the bottom surface of the mold, and the molten Mg is pushed up from the pouring port using a pump, a piston, a capillary phenomenon, or the like. In addition, prepare a mold and another container, connect the bottom of the mold and the bottom of the container with a connecting tube such as a hose, supply molten Mg to the container, and use the pressure applied to the molten Mg For example, the molten Mg may be supplied from the bottom surface side of the mold, or the molten Mg may be sucked from the opening side of the mold. In this form of supplying molten Mg from the bottom surface side of the mold, it is considered that unsolidified Mg or Mg alloy is supplied from the bottom surface side of the mold, so that the composite proceeds while compensating for the volume decrease. . Moreover, with this form, since the upper side is opening, it is anticipated that the air in a nonmetallic inorganic material will be easy to escape. And in this form, the composite member with few defects mentioned above can be produced by cooling an infiltration board to one direction toward the vertical direction lower side from the perpendicular direction upper side. However, the form using the dead weight of the molten Mg is considered to be superior in productivity.

<Raw material>
As the aggregate of non-metallic inorganic materials arranged in the mold, particulate or fibrous powder or a molded body can be used. When powder is used as a raw material, since it is excellent in fluidity, it can be easily filled into a mold or a mold and can be filled into a mold having a complicated shape. When the average particle diameter of the powder (in the case of a fiber, the average short diameter) is 1 μm or more and 3000 μm or less, particularly 10 μm or more and 200 μm or less, it is easy to fill a mold or the like, and the filling rate is easy to increase. When a plurality of types of powders having different average particle diameters are used in combination, the filling rate can be further increased. When using powder, for example, the mold is filled by tapping (giving vibration) to form an aggregate.

  On the other hand, as the compact, a powder compact obtained by solidifying the above powder with pressure or no pressure, a sintered compact obtained by sintering the powder compact, and other commercially available sintered compacts can be used. A composite member having a network portion can be manufactured by appropriately adjusting the sintering conditions. When using a molded body, (1) it is easier to obtain a composite member (substrate) in which the content of non-metallic inorganic material exceeds 70% by volume than when powder is used, and (2) strength that can be handled. (3) In the case of having a network part, the open pores are filled with Mg or Mg alloy so that the thermal conductivity is high and the thermal expansion coefficient is low. The effect that it is easy to obtain a substrate) is obtained.

  The powder compact can be formed by, for example, slip casting, pressure molding, doctor blade method, and the like. In slip casting, a slurry can be formed by using the above-described raw material powder, water, and a dispersant, and the slurry can be molded and then dried to form a powder compact. A general surfactant can be used as the dispersant. Examples of pressure molding include dry press, wet press, uniaxial pressure molding, CIP (hydrostatic press), and extrusion molding. In the case of dry press molding, the above-mentioned raw material powder is pressure-molded. In the case of wet press molding, a mixed powder obtained by mixing the raw material powder and a liquid such as water is pressure-molded to extrude the liquid. A powder molded body can be formed. The pressure during molding (molding pressure) may be appropriately selected. In the doctor blade method, a slurry is prepared using the above-mentioned raw material powder and a solvent, an antifoaming agent, a resin, etc., and this slurry is poured into a receptacle of a doctor blade to form a sheet-like body and then evaporate the solvent. A powder compact can be formed.

  Slip casting is a large molded body that can easily form a molded body with a complicated shape, and can obtain a molded body with a high filling rate (density) even when fine powder is used. Can be easily molded, and there is an advantage that there is little increase in equipment cost. In press molding, there are advantages that it is easy to make the particle size of the powder uniform, and that the number of steps is small and the productivity is excellent compared to slip casting. The doctor blade method can be suitably used when forming a plate-shaped molded body.

  When the powder compact is sintered, the sintering conditions can be appropriately selected according to the composition of the nonmetallic inorganic material. For example, in the case of SiC, the sintering conditions are (1) vacuum atmosphere, heating temperature: less than 800-1300 ° C, holding time: about 2 hours, or (2) air atmosphere, heating temperature: 800-1500 ° C, holding time : About 2 hours, or (3) vacuum atmosphere, heating temperature: 1300 ° C to 2500 ° C, holding time: about 2 hours. By performing the sintering, the strength is higher than that of the powder compact and easy to handle as described above, and a porous body can be easily produced. In addition, by adjusting the sintering temperature and holding time, the sintered body can be densified and the filling rate of non-metallic inorganic materials such as SiC can be improved, and the content of non-metallic inorganic materials is 70% by volume. It is easy to obtain the composite member as described above. Furthermore, by performing the sintering, it is possible to evaporate and remove the binder or the like used for the production of the powder compact by heating during the sintering. However, under the conditions (1) and (2), there is a tendency to obtain a composite member in a distributed form that does not have a network part. On the other hand, when sintering is performed under the condition (3), SiC can be directly bonded to each other. That is, the network part can be formed of SiC. By directly bonding SiC together, the strength of the sintered body becomes higher, and when this sintered body is used, the above-mentioned network part provides a composite with a low thermal expansion coefficient and high thermal conductivity. A member is easily obtained.

In addition, by forming a powder compact containing elements such as Si and sintering in a nitrogen atmosphere, a new non-metallic inorganic material (for example, Si ₃ N ₄ ) is generated, and this neoplasm forms a network. Forming a powder molded body using a plurality of types of powders having different compositions and forming a non-metallic inorganic material by reaction of some powders with molten Mg. Or a powder molded body impregnated with a solution of a precursor of a non-metallic inorganic material (e.g., polycarbosilane) and heated to produce a precursor-based non-metallic inorganic material, and this product produces a network. A part may be formed. Even when a molded body having these network parts is used, a composite member having a network part can be obtained.

  In particular, when producing a composite member (substrate) having a non-metallic inorganic material content of more than 70% by volume, it has open pores and few closed pores, specifically with respect to the total volume of the non-metallic inorganic material. It is preferable to form a porous body of 10% by volume or less, preferably 3% by volume or less. Such a porous body can sufficiently have a path for impregnating with molten Mg, and molten Mg can be infiltrated by capillary action even if the gap between the nonmetallic inorganic materials is very small. Therefore, it is easy to obtain a composite member in which the open pores are sufficiently filled with molten Mg and the pores are few.

<Oxide film formation>
When SiC is used as the non-metallic inorganic material, it is preferable to use a SiC aggregate provided with molten Mg having an oxide film on the surface because wettability between the aggregate and molten Mg is enhanced. By making the SiC aggregate including an oxide film, molten Mg easily penetrates due to capillary action even when the content of SiC is large and the gap between SiC is very small. When obtaining a composite member having a network part, it is preferable to include an oxidation step of forming an oxide film after producing a SiC aggregate such as a sintered body. Typically, when obtaining a composite member having no network part, When the SiC powder compact is subjected to molten Mg, an oxide film is formed on a raw material powder such as SiC powder, and a SiC aggregate (powder compact) is preferably formed using powder containing the oxide film.

  The conditions for forming the oxide film are the same for both SiC powder and sintered body, and the heating temperature is preferably 700 ° C or higher, particularly 750 ° C or higher, more preferably 800 ° C or higher, especially 850 ° C or higher. Further, 875 ° C. or higher and 1000 ° C. or lower is preferable. Further, it is preferable to form the oxide film so that the mass ratio of the raw material to SiC is 0.4% or more and 1.5% or less (the thickness of the oxide film: about 50 nm to 300 nm), particularly 1.0% or less. When an oxide film is formed, the oxygen concentration tends to be higher in the vicinity of SiC in the composite member (region within 100 to 300 nm from the outline of the SiC aggregate) than in locations other than the vicinity.

<Composite>
When the composite process of infiltrating molten Mg into the aggregate of nonmetallic inorganic materials arranged in the mold is performed in an atmosphere at atmospheric pressure (approximately 0.1 MPa (1 atm)) or less, it is difficult to take in the gas in the atmosphere, Stomach due to uptake is unlikely to occur. However, since Mg has a high vapor pressure, it becomes difficult to handle molten Mg in a high vacuum state. Therefore, 0.1 × 10 ⁻⁵ MPa or more is preferable when the atmospheric pressure in the composite process is less than atmospheric pressure. Further, when the composite process is performed in an inert atmosphere such as Ar, the reaction between the Mg component and the atmospheric gas can be prevented, and the deterioration of the thermal characteristics due to the presence of the reaction product can be suppressed. The infiltration temperature is preferably 650 ° C. or higher. Since the wettability increases as the infiltration temperature increases, 700 ° C. or higher, particularly 750 ° C. or higher is preferable. However, if the temperature exceeds 1000 ° C., defects such as shrinkage cavities and gas holes may occur, and Mg may boil. Therefore, the infiltration temperature is preferably 1000 ° C. or less. In particular, 900 ° C. or lower is preferable in order to suppress the formation of an excessive oxide film and the formation of crystallized substances.

<Cooling>
The most characteristic feature of the production method of the present invention is that the infiltration plate is cooled in one direction from the side opposite to the molten Mg supply side of the infiltration plate. For example, when the molten Mg supply side is set to the upper side in the vertical direction and the opposite side of the supply side is set to the lower side in the vertical direction, that is, when molten Mg is supplied in the direction of gravity, one direction from the lower side in the vertical direction to the upper side in the vertical direction Cool the infiltration plate. Alternatively, when the molten Mg supply side is set to the lower side in the vertical direction and the opposite side of the supply side is set to the upper side in the vertical direction, that is, when molten Mg is supplied in the direction opposite to the gravity direction, the upper side in the vertical direction is shifted to the lower side in the vertical direction. Cool in one direction, that is, along the direction of gravity. More specifically, for example, when manufacturing a rectangular plate-shaped composite member, one end side in the longitudinal direction of the infiltration plate is set as the lower side in the vertical direction, and the other end side is set as the upper side in the vertical direction. Cool toward the other end side (or from the other end side toward one end side). In order to cool the infiltrating plate in one direction in this way, for example, the side opposite to the molten Mg supply side (hereinafter referred to as the cooling start side) of the infiltrating plate may be forcibly cooled. The forced cooling includes liquid cooling using a liquid refrigerant, air cooling that performs forced air blowing, and the like. When using a coolant such as a liquid coolant, the coolant may be brought into direct contact with the bottom surface of the mold (or the mold opening) disposed on the cooling start side, or the bottom surface of the mold disposed on the cooling start side. The part (or the opening of the mold) may be disposed close to the refrigerant. In addition, a part other than the part where forced cooling is performed, such as the bottom part of the mold, may be covered with a heat insulating material, or the mold may be moved from the high temperature region to the low temperature region from the cooling start side. Any cooling method can be used, and the various cooling methods described above may be used in combination.

《Temperature gradient》
In the cooling step, when the infiltration plate is cooled such that the temperature gradient along the cooling direction of the infiltration plate (for example, the temperature gradient from the lower side in the vertical direction toward the upper side in the vertical direction) falls within the following specific range. The area ratio of the defective portion can be further reduced, and a higher-quality composite member with fewer defects can be easily obtained. Specifically, a temperature difference of a predetermined magnitude is provided between one side to be cooled first (for example, the lower side in the vertical direction) and the other side to be cooled later (for example, the upper side in the vertical direction). More specifically, a plurality of temperature measurement points are taken along the cooling direction of the infiltration plate, and when each temperature measurement point reaches a predetermined temperature, one temperature measurement adjacent to the temperature measurement point P _s A temperature difference ΔT _P between the point P _u and the other temperature measurement point P _d is taken, and this temperature difference ΔT _P is divided by the distance l between the two temperature measurement points P _u , P _d : ΔT _P / l Is the temperature gradient at the temperature measurement point. At this time, the infiltration plate is cooled so that the temperature gradient at each temperature measurement point is 0.01 ° C./mm or more. As the temperature gradient is increased, the area ratio of the defect portion tends to be reduced, and is preferably 0.1 ° C./mm or more, more preferably 0.5 ° C./mm or more.

  The temperature measurement points can be arbitrarily selected, but it is preferable to select a plurality of temperature measurement points at regular intervals with respect to the longitudinal direction of the infiltration plate. In addition, if the interval between adjacent temperature measurement points is too close, it is difficult to provide a sufficient temperature difference, and if it is too wide, the temperature difference becomes too large, which depends on the size of the infiltration plate in the longitudinal direction. 5 to 10 mm is considered preferable.

<Cooling rate>
In the cooling step, the speed at which the infiltration plate is cooled along the cooling direction (for example, the speed at which the cooling is performed from the lower side in the vertical direction toward the upper side in the vertical direction) is set to the following specific range. When cooled, the area ratio of the defective portion can be further reduced, and a higher-quality composite member with fewer defects can be easily obtained. Specifically, a plurality of temperature measurement points are taken along the cooling direction of the infiltration plate, and the time t required for each temperature measurement point to drop from a predetermined high temperature _TH to a predetermined low temperature T _L And the difference between the temperatures T _H and T _L : T _H −T _L divided by the time t is (T _H −T _L ) / t as the cooling rate at each temperature measurement point. At this time, cooling is performed so that the cooling rate at each temperature measurement point is 0.5 ° C./min or more. As the cooling rate is increased, the area ratio of the defective portion tends to be reduced, and is preferably 3 ° C./min or more, particularly 10 ° C./min or more, and more preferably 50 ° C./min or more. The faster the cooling rate, the finer the crystal grains of Mg or Mg alloy and the better the appearance of the composite member.

  The temperature gradient and the cooling rate can be changed by the temperature of the refrigerant, the amount of refrigerant (such as the amount of blown air), the distance between the refrigerant and the mold, the arrangement of the heat insulating material, the moving speed of the mold, and the like. When the temperature gradient is constant, the larger the cooling rate is, and when the cooling rate is constant, the larger the temperature gradient is, the easier it is to reduce the area ratio of the defective portion and the less deformation. Furthermore, if the cooling rate is high and the temperature gradient is large, it is easy to obtain a very high-quality composite member with very few defects.

  In realizing the above temperature gradient and cooling rate, if the temperature of molten Mg or Mg alloy is directly measured by a temperature measuring means such as a thermocouple, the temperature measuring means is significantly damaged. Therefore, it is advisable to prepare a mold that can realize the temperature gradient and the cooling rate and use it.

<Formation of metal coating layer>
Various methods can be used when manufacturing a composite member having a metal coating layer. For example, a metal coating layer may be separately formed after the substrate is formed. In this case, for example, an appropriate metal plate is prepared, and at least one method of brazing, ultrasonic bonding, cast-in, rolling (clad rolling), hot press, oxide solder method, bonding with an inorganic adhesive is used. Thus, a metal coating layer can be formed on the surface of the substrate. By using a metal plate, a metal coating layer having a composition different from that of the metal component in the substrate can be easily formed.

  Alternatively, a gap is provided between the aggregate of the non-metallic inorganic material and the mold so that the molten Mg flows into the gap during the above-described combination, and when the metal coating layer is formed simultaneously with the formation of the infiltration plate, Excellent composite member productivity. In this case, in the manufacturing method of the present invention in which cooling is performed in a specific direction as described above, a high-quality substrate can be formed, so that defects such as surface scratches can be reduced in the metal coating layer formed on the surface. Can do. In particular, even when the metal coating layer is thin, according to the production method of the present invention, a metal coating layer having excellent surface properties can be formed. In this embodiment, a metal coating layer having a thickness corresponding to the size of the gap can be formed. Therefore, the size of the gap may be adjusted so that the metal coating layer has a desired thickness.

  In particular, when a molded body of a nonmetallic inorganic material is used, the molded body is relatively excellent in strength and can be self-supporting in a mold. Therefore, for example, the gap is easily provided by forming a molded body smaller than the inner space of the mold and storing it in the mold. Spacers may be arranged so as to reliably maintain the gap. As the constituent material of the spacer, a material that can be removed by sublimation such as naphthalene, or a material having excellent heat resistance such as carbon, iron, and stainless steel (SUS430) can be used. In the latter case, the spacer may be left embedded in the metal coating layer, or the spacer portion may be removed by cutting or the like. Examples of the form of the spacer include a plate-like body and a linear body (wire). For example, a gap may be provided between the molded body and the mold by preparing a linear body having a slightly smaller diameter than the metal coating layer to be formed and fixing the molded body to the mold with the linear body. . In this case, since most of the linear body is embedded in the metal coating layer, a composite member having a good appearance can be obtained even if the linear body remains.

  The manufacturing method of the composite member of the present invention can manufacture a high-quality composite member even if it is large. Although the composite member of the present invention is large, defects are not concentrated and high quality. The heat radiating member of the present invention is excellent in thermal characteristics by being composed of the above-described composite member of the present invention. The semiconductor device of the present invention has excellent thermal characteristics by including the heat dissipation member.

FIG. 1 is an exploded perspective view showing an outline of a mold used for producing a composite member. 2 is a micrograph (50 ×) of a cross section of the composite member. FIG. 2 (I) shows the composite member of Embodiment 1, and FIG. 2 (II) shows the composite member of Embodiment 2. FIG. 3 is an explanatory view showing a state in which the mold used for producing the composite member is charged into the atmospheric furnace. FIG. 4 is an explanatory view showing a state in which the mold is cooled by water-cooled copper. FIG. 4 (I) shows an example in which no heat insulating material is disposed so as to contact the mold, and FIG. 4 (II) shows that the mold is insulated. An example covered with a material is shown. FIG. 5 is an explanatory view showing a state in which the mold is moved from the high temperature region to the low temperature region to be cooled.

(Embodiment 1)
Composite members (substrates) composed of pure magnesium and SiC were produced under various conditions, and the defect state and thermal characteristics of the obtained composite members were investigated.

  Pure magnesium ingot composed of Mg and impurities of 99.8% by mass or more as raw materials, and commercially available SiC sintered body (relative density 80%, length (long side) 200 mm x width (short side) 100 mm x thickness 5 mm) Prepared.

  The prepared SiC sintered body was oxidized at 875 ° C. for 2 hours to form an oxide film, and the wettability with molten pure magnesium was improved. The oxidation treatment step may be omitted.

  The SiC sintered body was accommodated in the mold 10 shown in FIG. 1, and molten pure magnesium was infiltrated into the sintered body 20 to produce an infiltration plate. In addition, the same code | symbol shows the same name thing in the following figures.

  The mold 10 is a vertically long rectangular parallelepiped box that is open on one side, and includes a main body 11 and a lid 12 as shown in FIG. The main body 11 includes a rectangular bottom surface portion 10b, a pair of rectangular end surface wall portions 10e standing from the bottom surface portion 10b, and side wall portions standing from the bottom surface portion 10b and connected to the pair of end surface wall portions 10e. With 10s. The side wall portion 10s has a vertically long rectangular shape as shown in FIG. The lid portion 12 is a vertically long rectangular plate that is equal to the side wall portion 10s, is disposed so as to face the side wall portion 10s, and is bolted (not shown) to the bottom surface portion 10b and the pair of end surface wall portions 10e. Fixed. By attaching the lid 12 to the main body 11, a vertically long mold 10 having an open side facing the bottom surface 10b is configured. Here, the mold 10 was made of carbon. The internal space of the mold 10 is used as a storage space for the sintered body 20. Here, the internal space of the mold 10 is sized according to the sintered body, and when the sintered body 20 is housed in the mold 10, a substantial gap is provided between the sintered body 20 and the mold 10. I tried not to. The lid 12 may be attached after the sintered body 20 is disposed on the main body 11, or the sintered body 20 may be disposed after the mold 10 is assembled. In addition, you may utilize the casting_mold | template integrally formed without setting it as the structure which combined the division | segmentation piece.

  Further, here, a commercially available mold release agent was applied to a portion of the inner peripheral surface of the mold 10 that contacts the sintered body, and then the sintered body 20 was accommodated in the mold 10. By applying the release agent, the composite member can be easily taken out. This step of applying the release agent may be omitted.

  The mold 10 has an ingot placing part (not shown) connected to the periphery of the opening. The said ingot prepared in the said ingot mounting part is arrange | positioned, and the said ingot is fuse | melted by heating this casting_mold | template 10 to predetermined temperature. The mold 10 is heated using an atmospheric furnace 30 shown in FIG. The atmosphere furnace 30 includes a sealable container 31, a heat insulating material 32 disposed on the inner peripheral surface of the container 31, and a heater 33 disposed in a space surrounded by the heat insulating material 32. The container 31 includes a bottom surface portion, a side wall portion erected from the bottom surface portion, and an upper surface portion that is disposed to face the bottom surface portion and is connected to the side wall portion. The heat insulating material 32 is disposed along the bottom surface, the side wall, and the top surface of the container 31. Here, the atmosphere furnace 30 is disposed so that the bottom surface portion is on the lower side in the vertical direction and the upper surface portion is on the upper side in the vertical direction. Here, the mold 10 is inserted so that the bottom surface portion 10b of the mold 10 is in contact with the bottom surface portion of the atmospheric furnace 30. That is, the mold 10 is loaded so that the bottom surface portion 10b of the mold 10 is on the lower side in the vertical direction and the opening of the mold 10 is on the upper side in the vertical direction.

  The atmosphere furnace 30 was adjusted so that the infiltration temperature was 775 ° C., the Ar atmosphere, and the atmospheric pressure: atmospheric pressure. Pure magnesium that is heated and melted by the heater 33 flows into the internal space of the mold 10 from the opening of the mold 10 due to its own weight, and is infiltrated into the sintered body 20 disposed in the internal space. A board is obtained.

  The mold 10 on which the infiltration plate was placed was cooled by the following cooling methods (1) to (7) to solidify pure magnesium, thereby obtaining a composite member having a length of 200 mm × width of 100 mm × thickness of 5 mm. Further, in the following cooling methods (2) to (7), the temperature gradient and the cooling rate described later were changed by appropriately adjusting the thickness of the heat insulating material, the blower amount of the fan, the temperature of water-cooled copper, and the like.

  (1) The heater 33 of the atmosphere furnace 30 is turned off, and the mold 10 is gradually cooled in the furnace 30. That is, the mold 10 is cooled in the state shown in FIG. In this cooling method, the infiltration plate in the mold is cooled so as to converge from the peripheral edge toward the central portion.

  (2) After turning off the heater 33 of the atmospheric furnace 30, the heat insulating material 32 on the bottom surface in the furnace 30 is removed so that the bottom surface 10b of the mold 10 is in direct contact with the container 31 of the furnace 30. Then, the mold 10 is cooled in the furnace 30.

  (3) After turning off the heater 33 of the atmospheric furnace 30, the heat insulating material 32 on the bottom surface in the furnace 30 is removed, and outside the furnace 30 and in the vicinity of the bottom surface of the furnace 30. A fan (not shown) is arranged, and the mold 10 is cooled in the furnace 30 while blowing air to the bottom surface of the furnace 30 by the fan. That is, the mold 10 is forcibly cooled by air cooling.

  (4) After turning off the power of the heater 33 of the atmospheric furnace, as shown in FIG. 4 (I), a part of the heat insulating material 32 on the bottom surface in the furnace is partially removed, and water-cooled copper 40 is attached. Water-cooled copper 40 is brought into contact with the bottom surface portion 10b, and the mold 10 is cooled in the furnace. That is, the mold 10 is forcibly cooled by liquid cooling. In FIG. 4, the container is omitted. Water-cooled copper 40 is commercially available.

  (5) After turning off the power to the heater 33 of the atmospheric furnace, as shown in FIG. 4 (II), the end wall of the mold 10, the side wall 10s, the lid 12, and the outer peripheral surface of the opening are covered with the heat insulating material 34. cover. That is, the heat insulating material 34 covers the part other than the bottom surface of the mold 10. In this state, the mold 10 is cooled in the same manner as the above (1) to (4). FIG. 4 (II) shows a state in which the mold 10 is covered with the heat insulating material 34 and the water-cooled copper 40 is in contact with the bottom surface portion 10b of the mold 10.

  (6) After turning off the heater 33 of the atmospheric furnace, the mold 10 is taken out of the furnace, and the water-cooled copper 40 is brought into contact with the bottom surface of the mold 10 to cool the mold 10. By taking out the mold 10 out of the atmosphere furnace and using the water-cooled copper 40, the cooling rate can be increased.

  (7) As shown in FIG. 5, in an atmosphere furnace (not shown), a high temperature region including a heater 33 and a heat insulating material 35 covering the outer periphery thereof, and a lower side of the high temperature region, water cooling An atmosphere furnace having a low temperature region made of copper 40 is prepared. The heater 33, the heat insulating material 35, and the water-cooled copper 40 are disposed so as to surround the periphery of the mold 10. The mold 10 has a bottom surface 10b supported by a bottomed cylindrical support base 50, and a movable part 51 that is movable in the vertical direction by a drive part (not shown) is disposed inside the support base 50. Yes. The support base 50 can be moved in the vertical direction by moving the movable portion 51 in the vertical direction. That is, the mold 10 can be moved in the vertical direction by the movable portion 51. Using such an atmospheric furnace, the movable part 51 is driven to cool the mold 10 by moving it from the high temperature region to the low temperature region. In other words, cooling is performed by entering the low temperature region from the bottom surface side of the mold 10 arranged on the lower side in the vertical direction.

  In cooling the infiltrating plate, the temperature gradient and the cooling rate were measured as follows for the cooling methods (1) to (7) described above. The results are shown in Table 1. And the cooling of (1)-(7) mentioned above was performed using the casting_mold | template 10 which achieves the temperature gradient shown in Table 1, and the cooling rate, the atmospheric furnace 30, etc., and the composite member was obtained.

  Thermocouples are arranged in the interior space of the vertically long mold 10 at equal intervals along the longitudinal direction from the bottom surface portion 10b side (vertical direction lower side) to the opening side (vertical direction upper side). Specifically, a thermocouple (not shown) is installed on the side wall 10s (or the lid 12) at an interval of 5 mm from the surface of the bottom surface 10b on the inner peripheral surface of the mold 10. The mold 10 is charged into an atmosphere furnace in an empty state in which pure magnesium and SiC are not filled. Thermocouples are commercially available.

<Temperature gradient>
The location where the thermocouple is placed is a temperature measurement point, and when the temperature at each temperature measurement point reaches 650 ° C., two temperature measurement points P _u and P _d adjacent to this temperature measurement point P _s , that is, the temperature temperature difference between the temperature T _d of the temperature T _u and the temperature measurement point P _d of the bottom surface side (the lower side in the vertical direction) of the temperature measuring point P _u of the opening side sandwiching the measurement point P _s (upward vertical direction): [Delta] T It takes _P = T _u -T _d, the temperature difference: T _u -T _d the two temperature measurement points P _u, the value obtained by dividing the distance l between P _d (where 10mm are): (T _u -T _d ) / l is the temperature gradient. Table 1 shows the minimum value among the temperature gradients obtained for the 40 temperature measurement points provided on the mold 10. In Table 1, samples with a temperature gradient value of `` minus '' are those on the infiltration plate housed in the mold, from the vertical upper side (mold opening side) to the vertical lower side (mold bottom side). ), Which corresponds to a sample that is convergently cooled from the periphery of the infiltration plate toward the central portion. A sample with a positive temperature gradient is a sample whose cooling proceeds from the lower side in the vertical direction (bottom side of the mold) to the upper side in the vertical direction (opening side of the mold) in the infiltration plate stored in the mold. Yes, here, it corresponds to a sample that is cooled in one direction from the lower side in the vertical direction to the upper side in the vertical direction on the infiltration plate.

<Cooling rate>
The location where the thermocouple is placed is a temperature measurement point, and the time t required for each temperature measurement point to drop from a high temperature T _H : 680 ° C to a low temperature T _L : 620 ° C is measured, and the temperature T _H , T _L difference: T _H −T _L (here 60 ° C.) divided by the above time t: (T _H −T _L ) / t is the cooling rate. Table 1 shows the minimum value among the cooling rates obtained for the 40 temperature measurement points provided in the mold 10.

  For the obtained composite member, the defect area ratio (%), dimensional difference (μm), surface roughness Ra (μm), thermal conductivity κ (W / m · K), thermal expansion coefficient α (ppm / K) ) Was measured. The results are shown in Table 1.

  The area ratio (%) of the defect portion was measured as follows. When the obtained composite member is viewed in plan, the composite member is cut by ion beam processing using a straight line passing through the center of gravity (here, the intersection of diagonal lines of a rectangular surface of 200 mm × 100 mm) as the thickness of the composite member. A cross section in the direction (here, a cross section having a cross-sectional area of 200 mm × 5 mm) is taken. In this cross section, an area in a range of up to 10% of the length of the cutting line (here, 200 mm) along the longitudinal direction of the cutting line with the center of gravity as the center is defined as a central area. Twenty small regions of 1 mm × 1 mm are arbitrarily selected from this central region (here, a cross-sectional area of 40 mm × 5 mm), and the ratio of the area of the defect portion to the area of each small region: the area ratio is obtained. The area of the defective portion is obtained using the cross-sectional image. Specifically, image analysis is performed on the image by a commercially available image processing apparatus, and a portion (mainly a void) other than pure magnesium and SiC constituting the composite member in each small region is a defective portion of the small region, The total area is obtained by the image processing apparatus. Table 1 shows the maximum value among the area ratios of 20 defect portions in each sample.

  As for the dimensional difference, the maximum thickness and the minimum thickness were measured in the image of the cross section (cross-sectional area: 200 mm × 5 mm) of each sample, and the difference between the maximum thickness and the minimum thickness was obtained. When this difference is 200 μm (0.2 mm) or more, it can be said that the surface is highly scratched and the surface properties are poor, and when it is 50 μm (0.05 mm) or less, the surface is very small.

  The surface roughness Ra was measured according to JIS B 0601 (2001).

  The thermal expansion coefficient α and the thermal conductivity κ were measured using a commercially available measuring instrument after cutting out a test piece from the obtained composite member. The thermal expansion coefficient α was measured in the range of 30 ° C to 150 ° C.

  Further, among the obtained composite members, the components of the defective portion having an area ratio of 10% or less were examined for the components of the composite member, the shape of SiC, the content of SiC, and the composite state. The components of the composite member were examined by an EDX apparatus, and were Mg and SiC, and the balance: inevitable impurities, which were the same as the raw materials used. Further, when the obtained composite member was subjected to CP (Cross-section Polisher) processing to obtain a cross section, and this cross section was examined by SEM observation, SiC was directly bonded to each other. That is, the network part was a porous body formed of SiC and was the same as the sintered material used. Furthermore, when the cross sections of these samples were observed with an optical microscope (50 ×), it was confirmed that pure magnesium was infiltrated into the gaps between SiC as shown in FIG. 2 (I). In FIG. 2, the portion constituting the continuous network is SiC, and the portion solidified in a granular form is pure magnesium.

  The content of the SiC is observed with an optical microscope (50 times) any cross section of the composite member, this observation image is image processed with a commercially available image analysis device, to determine the total area of SiC in the cross section, A value obtained by converting the total area into a volume ratio was defined as a volume ratio based on the cross section, and a volume ratio of the cross section of n = 3 was obtained and averaged (area ratio≈volume ratio). As a result, the content of SiC was 80% by volume.

  As shown in Table 1, a composite member obtained by cooling the infiltration plate in one direction from the opposite side of the molten magnesium supply side in the infiltration plate, here, from the lower vertical direction to the upper vertical direction. The composite member obtained by cooling the infiltrated plate in one direction is a large plate that can take a circular region with a diameter of more than 50 mm. It can be seen that the quality. In particular, any defects present in the composite member that had been cooled in one direction were so fine that it was difficult to visually confirm (0.1 mm or less), and had a uniform structure. Moreover, it turns out that the composite member obtained by performing the said one-way cooling has few surface defects, and is excellent in a dimensional accuracy. Furthermore, the composite material obtained by performing the above-mentioned one-way cooling has a very low coefficient of thermal expansion of about 4 ppm / K, and is compatible with semiconductor elements having a coefficient of thermal expansion of about 4 ppm / K and its peripheral components. In addition to being excellent, there are samples with a thermal conductivity of 180 W / K · m or more, especially 300 W / K · m or more, and it can be seen that they have excellent thermal conductivity.

  On the other hand, it can be seen that the composite member that has not been cooled in one direction has a large area ratio of the above-described defect portions and locally has defects (pores). Moreover, the defect which existed in this composite member was so large that it could be confirmed visually. Due to the presence of such large defects, it is considered that this composite member has low dimensional accuracy and thermal conductivity.

  Therefore, it is expected that the composite member obtained by performing the unidirectional cooling can be suitably used as a constituent material of the heat dissipation member of the semiconductor element.

  In addition, as shown in Table 1, by cooling so that the temperature gradient at each temperature measurement point is 0.01 ° C / mm or more and the cooling rate at each temperature measurement point is 0.5 ° C / min or more, the defective portion It can be seen that the area ratio can be reduced. Furthermore, it can be seen that the larger the temperature gradient and the cooling rate, the smaller the area ratio of the defect portion. In addition, as shown in Table 1, it can be seen that when the cooling rate is constant, the defect portion can be reduced as the temperature gradient increases. From this, it can be said that it is difficult to obtain a composite member having defects locally by performing the cooling in one direction and increasing the temperature gradient and the cooling rate.

  In the first embodiment, a commercially available SiC sintered body is used, but a composite member can be manufactured using SiC powder. Specifically, for example, the SiC powder is subjected to an oxidation treatment at 875 ° C. for 2 hours, and then a powder compact is formed on the mold using tapping, slip casting, etc., and melted in the same manner as in the first embodiment. By infiltrating the pure magnesium into the powder compact and controlling the cooling conditions during solidification in the same manner as in the first embodiment, a composite member having a large size and few defects can be obtained as in the first embodiment. The powder compact may be appropriately sintered.

(Embodiment 2)
A composite member comprising a substrate made of a composite material made of a composite of pure magnesium and SiC and a metal coating layer covering each of the two opposing surfaces of the substrate was produced, and the defect state and thermal characteristics of the resulting composite member were measured. Examined.

  As raw materials, the same pure magnesium ingot as in Embodiment 1 and a SiC sintered body were prepared. The SiC sintered body was subjected to the same oxidation treatment as in the first embodiment. Further, a pair of carbon plate-like spacers having a length of 10 mm, a width of 100 mm, and a thickness of 0.5 mm were prepared.

  Here, the mold 10 shown in FIG. 1 used in the first embodiment and having a size that allows the spacer to be disposed between the SiC sintered body and the mold is used. A sintered body 20 and a pair of spacers (not shown) are accommodated in a mold 10 appropriately coated with a release agent, and the sintered body 20 is sandwiched between the pair of spacers. By being sandwiched between the spacers, the sintered body 20 is stably disposed in the mold, and between the surface of the sintered body 20 and the side wall portion 10s of the mold 10, and the back surface of the sintered body 20 and the mold. A gap corresponding to the thickness of the spacer (here, 0.5 mm) is provided between each of the 10 lid portions 12. This mold 10 was charged into an atmospheric furnace as in the first embodiment. Then, under the same conditions as in Embodiment 1, the sintered body 20 and molten pure magnesium were combined to produce an infiltration plate. In this embodiment, at the same time as forming the infiltration plate, the melted pure magnesium flows into the gap between the sintered body provided by the spacer and the mold as described above, so that the two opposite surfaces of the infiltration plate are opposed to each other. Each layer is made of pure magnesium.

  Here, using the water-cooled copper or the like so as to achieve the same temperature gradient and cooling rate as the sample No. 38 of Embodiment 1, the infiltration plate is moved in one direction from the vertical lower side to the vertical upper side. And pure magnesium was solidified. Through the above process, a composite member (length 200 mm × width 100 mm × thickness 6 mm) was obtained.

  When the cross section of the obtained composite member was observed with an optical microscope (50 ×), as shown in FIG. 2 (II), a substrate made of a composite material in which pure magnesium was infiltrated into the gaps between SiC, and the substrate It was confirmed that the surface was provided with a metal coating layer made of pure magnesium. When the composition of the constituent metals of the substrate and the metal coating layer was examined with an EDX apparatus, the same composition (pure magnesium) was obtained. Further, from the observation image of the cross section, it was confirmed that each metal coating layer had a structure continuous with pure magnesium in the substrate. Furthermore, when the thickness of each metal coating layer was measured using the observation image of the cross section, it was about 0.5 mm (500 μm), and it was confirmed that it substantially coincided with the thickness of the spacer.

  In the obtained composite member, the content of SiC in the portion where pure magnesium and SiC were combined, that is, the portion excluding the metal coating layer was measured and found to be 80% by volume. The SiC content was measured in the same manner as in the first embodiment.

  About the obtained composite member, the area ratio (%), the dimensional difference (μm), and the surface roughness Ra (μm) of the defect portion were measured in the same manner as in Embodiment 1. The area ratio: 0.11%, the dimensional difference 1 μm and surface roughness Ra: 0.8 μm. In addition, the dimensional difference measured thickness including the metal coating layer. Moreover, the surface roughness Ra measured the surface of the metal coating layer.

  When the thermal expansion coefficient α (ppm / K) and the thermal conductivity κ (W / mK) were measured in the same manner as in Embodiment 1 for the obtained composite member, the thermal expansion coefficient α: 5.1 ppm / K, The thermal conductivity κ was 250 W / m · K.

  From the above, even in the case of a composite member that is large enough to have a circular area with a diameter of more than 50 mm and that has a metal coating layer, the infiltrating plate is lowered in the vertical direction as in the first embodiment. It can be seen that by cooling in one direction from the side toward the upper side in the vertical direction, there is no large defect locally, the surface property is excellent, and a high-quality composite member can be obtained. In addition, this composite member is excellent in compatibility with a semiconductor element having a thermal expansion coefficient of about 4 ppm / K and its peripheral components, and also has excellent heat dissipation, and is suitable as a constituent material of the heat dissipation member of the semiconductor element. Expected to be available. In particular, the composite member of Embodiment 2 can be provided with Ni plating or the like by electroplating by providing metal coating layers on both sides of the substrate. By applying Ni plating or the like, the wettability with the solder can be improved, and even when it is used in a semiconductor device where solder application is desired, it can sufficiently cope. In addition, since the surface of the substrate itself is smooth as described above, the surface of the metal coating layer is also smooth, and the plating can be formed with a uniform thickness.

  In the composite member of Embodiment 2, the thickness and formation region of the metal coating layer can be easily changed by appropriately selecting the thickness and shape of the plate-like spacer and the number used. For example, by disposing the spacer only on one surface of the formed body (sintered body), a composite member having a metal coating layer only on one surface of the substrate can be obtained.

  The present invention is not limited to the above-described embodiment, and can be appropriately changed without departing from the gist of the present invention. For example, composition of non-metallic inorganic material in composite member, content, presence / absence of network part, constituent material of network part, composition of metal component (for example, magnesium alloy), size of composite member, thickness of metal coating layer The conditions at the time of combining the composite members can be changed as appropriate.

  The composite member of the present invention can be suitably used for a heat spreader of the semiconductor element (the heat dissipation member of the present invention). The manufacturing method of this invention composite member can be utilized suitably for manufacture of the said composite member. The semiconductor device of the present invention can be suitably used for constituent members of various electronic components.

10 Mold 10b Bottom 10e End wall 10s Side wall 11 Body
12 Lid 20 Sintered body (molded body) 30 Atmospheric furnace 31 Container
32, 34, 35 Insulation 33 Heater 40 Water-cooled copper 50 Support base 51 Moving parts

Claims (18)

A method for producing a composite member in which magnesium or a magnesium alloy and a nonmetallic inorganic material are combined,
A combined process of infiltrating molten magnesium or a magnesium alloy into an aggregate of nonmetallic inorganic materials housed in a mold to form an infiltrated plate;
Wherein said infiltrant plate is cooled from the side opposite the supply side of the molten magnesium or magnesium alloy in one direction in the infiltrated plate, immediately give a cooling step of solidifying the molten magnesium or magnesium alloy,
The non-metallic inorganic material aggregate includes non-metallic inorganic material powder filled in the mold, non-metallic inorganic material powder compacted by pressing or non-pressurizing, and non-metallic inorganic material method of producing a composite member according to claim 1 Tanedea Rukoto selected from the sintered body. In the composite step, the molten magnesium or magnesium alloy is supplied from the upper side in the vertical direction toward the lower side in the vertical direction by its own weight with respect to the aggregate.
The cooling step includes cooling the infiltration plate in one direction from a lower vertical direction to an upper vertical direction in the infiltration plate to solidify the molten magnesium or magnesium alloy. A method for producing the composite member according to 1. In the cooling step, a plurality of temperature measurement points are taken along the cooling direction of the infiltration plate, and when each temperature measurement point reaches a predetermined temperature, one temperature measurement point adjacent to the temperature measurement point P _s The temperature difference between P _u and the other temperature measurement point P _d is taken, and the value obtained by dividing this temperature difference by the distance l between the two temperature measurement points P _u and P _{d is} the temperature gradient of the temperature measurement point P _s . 3. The method for manufacturing a composite member according to claim 1, wherein the infiltration plate is cooled so that a temperature gradient at each temperature measurement point is 0.01 ° C./mm or more. In the cooling step, a plurality of temperature measurement points are taken along the cooling direction of the infiltration plate, and the time t required for each temperature measurement point to drop from a predetermined high temperature T _H to a predetermined low temperature T _L And the difference between the temperatures T _H and T _L : T _H -T _L divided by the time t is the cooling rate at each temperature measurement point, the cooling rate at each temperature measurement point is 0.5 4. The method for producing a composite member according to claim 1, wherein the infiltrating plate is cooled so as to be equal to or higher than C / min.   5. The cooling process according to any one of claims 1 to 4, wherein in the cooling step, cooling is performed in one direction by forcibly cooling a side opposite to a supply side of the molten magnesium or magnesium alloy in the infiltration plate. The manufacturing method of the composite member of description.   6. The method for manufacturing a composite member according to claim 1, wherein the infiltration plate has a size capable of taking a circular region having a diameter of more than 50 mm. Ri composite member der produced by the production method of a composite member according to claim 1,
It is a plate material having a size capable of taking a circular region having a diameter of more than 50 mm,
Taking a straight line passing through the center of gravity when the composite member is viewed in plan as a cutting line, a cross section in the thickness direction of the composite member is taken, and in this cross section, up to 10% of the length of the cutting line centering on the center of gravity When the area is a central area, an arbitrary small area of 1 mm x 1 mm is taken from this central area, and the area ratio is the ratio of the area of the defect portion to the area of the small area, the area ratio is 10% or less A composite member characterized by   8. The composite member according to claim 7, wherein a difference between the maximum thickness and the minimum thickness of the composite member in the cross section is 0.2 mm or less. Manufactured by the method for producing a composite member according to any one of claims 1 to 6,
A composite member characterized by being a plate material having a size capable of taking a circular region having a diameter of more than 50 mm.   10. The composite member according to claim 7, wherein a surface roughness Ra of the composite member is 2.5 μm or less.   The composite member according to any one of claims 7 to 10, wherein the non-metallic inorganic material includes SiC.   12. The composite member according to claim 7, wherein the content of the nonmetallic inorganic material is 20% by volume or more.   The composite member according to any one of claims 7 to 12, wherein the composite member has a thermal conductivity of 180 W / m · K or more.   14. The composite member according to claim 7, wherein a thermal expansion coefficient of the composite member is 3.5 ppm / K or more and 20 ppm / K or less.   15. The composite member according to claim 7, wherein the composite member has a rectangular shape with a short side exceeding 50 mm and a long side exceeding 100 mm. The composite member is a substrate made of a composite material in which the magnesium or the magnesium alloy and the non-metallic inorganic material are combined;
16. The composite member according to claim 7, further comprising a metal coating layer covering at least one surface of the substrate.   A heat radiating member comprising the composite member according to any one of claims 7 to 16.   18. A semiconductor device comprising: the heat dissipating member according to claim 17; and a semiconductor element mounted on the heat dissipating member.