Copper molybdenum composite, heat dissipation plate, semiconductor device, method for manufacturing copper molybdenum composite, and method for manufacturing heat dissipation plate

Abstract

Provided is a copper molybdenum composite, which has a plate-shaped body having a first surface and a second surface, and in which the copper content is 20-80 mass%, inclusive, and the standard deviation of the in-plane distribution of the nickel content measured at a plurality of locations on the first surface and the second surface is 5 ppm or less.

Description

Copper-molybdenum composite, heat sink, semiconductor device, method for manufacturing copper-molybdenum composite, and method for manufacturing heat sink

[0001] The present disclosure relates to a copper-molybdenum composite, a heat sink, a semiconductor device, a method for manufacturing a copper-molybdenum composite, and a method for manufacturing a heat sink. This application claims priority to Japanese Patent Application No. 2024-055988 filed on March 29, 2024, the entire contents of which are incorporated herein by reference.

[0002] Patent Document 1 discloses a heat dissipation substrate and its manufacturing method that has a thermal expansion coefficient close to that of semiconductors and ceramic insulators, has a large area with excellent thermal conductivity, and can reduce costs due to the large number of processes and complexity. The heat dissipation substrate is made of a Cu-Mo composite substrate manufactured by impregnating a compact of molybdenum (Mo) with 20 to 60% copper (Cu) by mass.

[0003] Patent Document 2 discloses a copper-molybdenum alloy for use in heat sinks and the like, which is homogeneous, high-density, and has excellent thermal conductivity and bending strength, and a method for producing the alloy. The production method involves mixing copper powder and molybdenum powder with a trace amount of phosphorus, press-molding the resulting mixed powder into a predetermined shape, and then sintering the resulting compact to obtain a copper-molybdenum alloy sintered body containing 0.003 to 0.06 wt. % phosphorus. It is considered more preferable to add 0.1 to 0.5 wt. % of powder of one or more of cobalt, iron, and nickel in addition to the copper and molybdenum powders and phosphorus.

[0004] JP-A No. 11-307701 JP-A No. 8-253833

[0005] The copper-molybdenum composite of the present disclosure is a plate-like body having a first surface and a second surface, and has a copper content of 20 mass % or more and 80 mass % or less, and the standard deviation of the in-plane distribution of the nickel content measured at multiple locations on the first surface and the second surface is 5 ppm or less.

[0006] FIG. 1 is a schematic perspective view showing an example of a copper-molybdenum composite and a heat sink. FIG. 2 is a schematic cross-sectional view showing an example of a copper-molybdenum composite and a heat sink. FIG. 3 is a plan view of a copper-molybdenum composite illustrating a method for measuring nickel content. FIG. 4A is a schematic cross-sectional view showing an example of a heat sink. FIG. 4B is a schematic cross-sectional view showing an example of a heat sink. FIG. 5 is a schematic cross-sectional view showing an example of a heat sink having a metal coating on an end surface in the thickness direction. FIG. 6A is an exploded perspective view showing an example of a semiconductor device. FIG. 6B is a schematic cross-sectional view of the semiconductor device shown in FIG. 6A. FIG. 7 is a diagram illustrating heat treatment in a batch furnace, which is a conventional heat treatment method. FIG. 8 is a diagram illustrating heat treatment in a belt furnace, which is a heat treatment method according to the present disclosure. FIG. 9 is a flow chart illustrating a method for producing a copper-molybdenum composite according to an embodiment. FIG. 10 is a flow chart illustrating a method for producing a heat sink according to an embodiment. FIG. 11 is a flow chart illustrating a method for producing a heat sink according to another embodiment.

[0007] [Problem to be Solved by the Present Disclosure] Copper-molybdenum composite materials are characterized by having both high thermal conductivity and a low coefficient of linear expansion, and are used as heat sinks in applications such as semiconductor devices. One of the objects of the present disclosure is to provide a copper-molybdenum composite that, when used as a heat sink, has a stable high thermal conductivity and a coefficient of linear expansion close to that of semiconductors and ceramic insulators, and a heat sink using the same.

[0008] [Advantages of the Present Disclosure] The copper molybdenum composite of the present disclosure has a stable high thermal conductivity and a linear expansion coefficient close to that of semiconductors and ceramic insulators.

[0009] [Description of Embodiments of the Present Disclosure] First, embodiments of the present disclosure will be listed and described.

[0010] The inventors of the present disclosure discovered a problem in that the nickel content of conventional copper-molybdenum composites varies within the plane. The inventors identified the cause of this variation and came up with the idea of ​​using a new manufacturing method to obtain a copper-molybdenum composite with a stable in-plane distribution of nickel content.

[0011] (1) A copper-molybdenum composite according to an embodiment of the present disclosure is a plate-shaped body having a first surface and a second surface. The copper content of the copper-molybdenum composite is 20 mass % or more and 80 mass % or less. The standard deviation of the in-plane distribution of the nickel content measured at multiple locations on the first surface and the second surface is 5 ppm or less.

[0012] The copper-molybdenum composite has a stable in-plane thermal conductivity because the nickel content is uniform within the plane. There are no locally low thermal conductivity areas within the plane, making it less likely to develop heat spots when used for heat dissipation purposes. Furthermore, the copper-molybdenum composite also has a desired linear expansion coefficient depending on the ratio of copper to molybdenum. That is, the copper-molybdenum composite has a stable, high thermal conductivity and a linear expansion coefficient close to that of semiconductors and ceramic insulators.

[0013] The standard deviation is the standard deviation of a plurality of nickel content measurements obtained by inductively coupled plasma atomic emission spectrometry at a plurality of measurement points on the first and second surfaces, the plurality of measurement points including one or more measurement points in the central regions of the first and second surfaces and a plurality of measurement points in regions other than the central regions.

[0014] (2) In the above (1), the maximum nickel content may be 20 ppm or less.

[0015] Nickel is a metal with lower thermal conductivity than copper and molybdenum. Therefore, the copper-molybdenum composite has lower thermal conductivity as a composite material due to the inclusion of nickel. By keeping the nickel content below a certain level, the composite can have a stable, high in-plane thermal conductivity.

[0016] (3) A heat sink according to an embodiment of the present disclosure is a heat sink made of the copper-molybdenum composite of (1) or (2). The heat sink has a thermal conductivity in the thickness direction at room temperature of 150 W / m·K or more and 320 W / m·K or less. The heat sink also has a linear expansion coefficient in the in-plane direction perpendicular to the thickness direction from room temperature to 800°C of 6.0 ppm / K or more and 14.0 ppm / K or less.

[0017] A heat sink having the above characteristics has a stable, high thermal conductivity and a linear expansion coefficient close to that of semiconductors and ceramic insulators. Here, "room temperature" means 27°C unless otherwise specified. The thickness direction is the direction from the first surface to the second surface.

[0018] (4) The heat sink of (3) above may have a metal coating on at least a portion of the first surface or the second surface, and the metal coating may be made of at least one metal selected from the group consisting of nickel, nickel-phosphorus, copper, silver, gold, and platinum, or an alloy thereof.

[0019] A heat sink plate provided with a metal coating has good bonding properties with a bonding material used to bond the heat sink plate to a heat dissipation target.

[0020] (5) A heat sink according to another embodiment of the present disclosure has a laminated structure of at least one first material layer made of the copper-molybdenum composite described above in (1) or (2) and a second material layer containing copper as a main component.

[0021] A heat sink with multiple layers laminated together can easily achieve the desired thermal conductivity and linear expansion coefficient by changing the thickness of each layer and the number of layers, resulting in a heat sink that has a stable, high thermal conductivity and a linear expansion coefficient close to that of semiconductors and ceramic insulators.

[0022] (6) In the above (5), the heat sink may have a thermal conductivity in the thickness direction at room temperature of 180 W / m K or more and 350 W / m K or less. The heat sink may also have a linear expansion coefficient in an in-plane direction perpendicular to the thickness direction from room temperature to 800°C of 6.5 ppm / K or more and 14.0 ppm / K or less.

[0023] The second material layer, which is primarily composed of copper, has a higher thermal conductivity than the first material layer. By forming a laminated structure of the first material layer and the second material layer, a heat sink with even better thermal conductivity can be obtained.

[0024] (7) In the above (5) or (6), the heat sink may have upper and lower surfaces, which are end surfaces in the thickness direction, that are both layers made of the second material layer.

[0025] By using copper as the second material layer on the top and bottom surfaces, it is possible to improve the heat spread at the contact surface and the heat conduction performance in the thickness direction, resulting in a heat sink with a linear expansion coefficient close to that of semiconductors and ceramic insulators while also having a stable and high thermal conductivity.

[0026] (8) In any one of the above (5) to (7), the heat sink may have a metal coating on at least a portion of the upper or lower surface, which is an end surface in the thickness direction, and the metal coating may be made of at least one metal selected from the group consisting of nickel, nickel-phosphorus, copper, silver, gold, and platinum, or an alloy thereof.

[0027] A heat sink plate provided with a metal coating has good bonding properties with a bonding material used to bond the heat sink plate to a heat dissipation target.

[0028] (9) A semiconductor device according to an embodiment of the present disclosure includes a semiconductor element and the heat sink according to any one of (3) to (8) above. The semiconductor element and the heat sink are joined together by a bonding material, which is a solder material containing lead, a lead-free solder, a brazing material, a silver sintered material, or a copper sintered material.

[0029] In the semiconductor device having the above configuration, by using the heat sink of the present disclosure, thermal stress occurring at the interface between the semiconductor element or the insulator and the heat sink tends to be reduced.

[0030] (10) A method for producing a copper-molybdenum composite according to an embodiment of the present disclosure includes the steps of: preparing a composite material by either a method of infiltrating a molybdenum powder compact with copper or a method of infiltrating a powder compact made of a copper and molybdenum mixed powder with copper; and heat treating the composite material to produce a copper-molybdenum composite. The heat treatment is a process in which the composite material is passed through a heat treatment furnace while being fed at a feed rate of 100 mm / min or less. The interior of the heat treatment furnace is a reducing atmosphere. The composite material is held in the heat treatment furnace at a temperature of 700°C to 1000°C for 0.3 hours or more.

[0031] The above temperature and time are processing conditions for annealing the composite material. The inventors of the present disclosure discovered that nickel, contained as an impurity in molybdenum, tends to concentrate near the center of the plate-shaped body during the process of copper solidifying from a molten state. They then came up with the idea that the above heat treatment conditions can prevent this concentration, i.e., make the nickel dispersed. Therefore, the above manufacturing method can produce a copper-molybdenum composite that has a stable, high thermal conductivity and a linear expansion coefficient similar to that of semiconductors and ceramic insulators.

[0032] (11) In the above (10), the copper-molybdenum composite may be a plate-like body having a first surface and a second surface. The copper-molybdenum composite may have a copper content of 20 mass % or more and 80 mass % or less. The copper-molybdenum composite may have an in-plane distribution of nickel content measured at multiple locations on the first surface and the second surface with a standard deviation of 5 ppm or less.

[0033] The copper-molybdenum composite has a stable in-plane thermal conductivity because the nickel content is uniform within the plane. There are no areas with locally low thermal conductivity within the plane, making it less likely to develop heat spots when used for heat dissipation purposes. Furthermore, the copper-molybdenum composite has a desired linear expansion coefficient depending on the ratio of copper to molybdenum. That is, the copper-molybdenum composite has a stable, high thermal conductivity while also having a linear expansion coefficient close to that of semiconductors and ceramic insulators.

[0034] (12) In the above (11), the maximum value of the nickel content may be 20 ppm or less.

[0035] By keeping the nickel content at a certain level or less, it is possible to obtain a copper-molybdenum composite having a stable, high in-plane thermal conductivity.

[0036] (13) In any one of the above (10) to (12), the copper-molybdenum composite may be a plate-like body having a long side of 100 mm or more, a short side of 100 mm or more, and a thickness of 2 mm or more.

[0037] According to the above manufacturing method, it is possible to obtain a copper-molybdenum composite having a stable and high in-plane thermal conductivity, even when the copper-molybdenum composite has the above-mentioned size.

[0038] (14) A method for manufacturing a heat sink according to an embodiment of the present disclosure includes the steps of preparing the copper-molybdenum composite by any one of the manufacturing methods (10) to (13) above, and rolling the copper-molybdenum composite at a working ratio of 20% or more to form a heat sink. The heat sink may have a thermal conductivity in the thickness direction at room temperature of 150 W / m·K or more and 320 W / m·K or less. The heat sink may have a linear expansion coefficient from room temperature to 800°C in an in-plane direction perpendicular to the thickness direction of 6.5 ppm / K or more and 14.0 ppm / K or less.

[0039] The copper-molybdenum composite can be rolled to a desired thickness. The heat sink obtained by rolling the copper-molybdenum composite to a rolling processing ratio of 20% or more has a linear expansion coefficient close to that of semiconductors and ceramic insulators while maintaining a stable and high thermal conductivity as described above.

[0040] (15) In the above (14), a step of forming a metal coating on at least a portion of the first surface or the second surface, or on both the first surface and the second surface, of the heat sink may be further included. The metal coating may be at least one metal selected from the group consisting of nickel, nickel-phosphorus, copper, silver, gold, and platinum, or an alloy thereof.

[0041] A heat sink plate provided with a metal coating has good bonding properties with a bonding material used to bond the heat sink plate to a heat dissipation target.

[0042] (16) A method for manufacturing a heat sink according to another embodiment of the present disclosure includes a step of alternately stacking in the thickness direction at least one first material layer made of a copper-molybdenum composite obtained by any one of the manufacturing methods (10) to (13) above and a second material layer containing copper as a main component.

[0043] By using a heat sink composed of multiple layers, it is easy to obtain the desired thermal conductivity and linear expansion coefficient by changing the thickness and number of each layer, which makes it possible to obtain a heat sink having a stable high thermal conductivity and a linear expansion coefficient close to that of semiconductors and ceramic insulators.

[0044] (17) In the above (14) to (16), a step of cutting into small pieces with long sides less than 100 mm and short sides less than 100 mm may be further included.

[0045] The heat sink obtained by the above manufacturing method has a stable and high thermal conductivity within its surface. Therefore, the heat sinks obtained by cutting into smaller pieces also have a uniform thermal conductivity. Since each of the multiple small pieces of the heat sink has a uniform thermal conductivity, the manufacturing yield is improved.

[0046] [Details of the embodiments of the present disclosure] Specific examples of the copper-molybdenum composite, heat sink, and semiconductor device of the present disclosure are described below. The same reference numerals in the drawings indicate the same items. The sizes of the components shown in each drawing are expressed for the purpose of clarifying the description and do not necessarily represent the actual dimensional relationships.

[0047] <Copper-molybdenum composite> FIG. 1 is a schematic diagram showing a copper-molybdenum composite 10 according to an embodiment of the present disclosure. The copper-molybdenum composite 10 of the present disclosure is a plate-like body having a first surface 10a and a second surface 10b. The copper content is 20% by mass or more and 80% by mass or less, with the entire copper-molybdenum composite being 100% by mass. Furthermore, the standard deviation of the in-plane distribution of the nickel content measured at multiple locations on the first surface 10a and the second surface 10b is 5 ppm or less. While FIG. 1 illustrates an example of the shape of the copper-molybdenum composite 10 as a rectangular plate in plan view, the shape is not limited thereto. For example, the copper-molybdenum composite 10 may be a polygonal or circular plate in plan view, or may even be an irregularly shaped plate.

[0048] FIG. 2 is a diagram showing an example of a cross section perpendicular to the first surface 10a of the copper-molybdenum composite 10 of FIG. 1. The copper-molybdenum composite 10 has a structure in which copper is impregnated into a molybdenum powder compact. The ratio of copper to molybdenum is approximately uniform in the cross section shown in FIG. 2. Copper is a material with high thermal conductivity. Molybdenum is a material with a lower linear expansion coefficient than copper. The copper-molybdenum composite 10 contains 20% by mass or more and 80% by mass or less of copper, and thereby has a thermal conductivity and linear expansion coefficient intermediate between those of copper and molybdenum. The ratio of copper to molybdenum is determined by inductively coupled plasma optical emission spectroscopy (ICP-OES).

[0049] A standard deviation of the in-plane distribution of the nickel content of 5 ppm or less indicates that nickel is uniformly dispersed throughout the copper-molybdenum composite 10. In other words, it means that there are no areas where nickel is unevenly present. This characteristic is due to the manufacturing method of the copper-molybdenum composite described below. The standard deviation of the in-plane distribution of the nickel content may be 3 ppm or less, or may be 2 ppm or less. The lower limit of the standard deviation of the in-plane distribution of the nickel content is not limited, but may be 1 ppm or more. The copper-molybdenum composite 10 of the present disclosure has a uniform in-plane nickel concentration, and therefore a uniform in-plane thermal conductivity.

[0050] Nickel is contained as an inevitable impurity in the raw materials of the copper-molybdenum composite 10. Nickel has a lower thermal conductivity than copper and molybdenum. If the nickel concentration is high in a portion of the plate, that portion will have locally low thermal conductivity. When a copper-molybdenum composite 10 with a non-uniform nickel concentration is used as a heat sink, the heat dissipation characteristics between the first surface 10a and the second surface 10b will be poor in the locally low thermal conductivity portion. As a result, heat spots are likely to occur within the surface. Furthermore, if such a copper-molybdenum composite 10 is divided into multiple small heat sinks, the characteristics of the individual heat sinks may be non-uniform. In other words, the product yield is likely to be poor.

[0051] If the maximum nickel content measured in the plane is 20 ppm or less, the decrease in thermal conductivity due to nickel can be reduced. The copper-molybdenum composite 10 can have high thermal conductivity and uniform properties in the plane. The maximum nickel content may be 10 ppm or less, or 7 ppm or less. There is no particular limitation on the minimum nickel content, but as described above, nickel is contained in the material as an unavoidable impurity. Therefore, the minimum value may be 2 ppm or more, or 5 ppm or more.

[0052] (Method for Measuring Nickel Content) The nickel content is determined by ICP-OES. A method for determining the in-plane distribution of nickel content will be described with reference to FIG. 3. FIG. 3 is a diagram illustrating the method for measuring the nickel content. FIG. 3 shows a plan view of the first surface 10a of the copper-molybdenum composite 10. The second surface 10b is not shown, but the measurement method is the same. In FIG. 3, the copper-molybdenum composite 10 is rectangular, with one side having a length La and the other side having a length Lb. The direction of the side with length La corresponds to the belt traveling direction when the copper-molybdenum composite is processed using a belt furnace. The direction of the side with length La also corresponds to the rolling direction. The copper-molybdenum composite 10 is rolled in the direction from P6 to P7 in FIG. 3, that is, in the order of P6, P5, P8, and P7. Eight locations are selected within the plane to measure the nickel content. Measurement locations P1 to P4 are located at the four corners, each a length L1 inward from the outer edge of each side. Measurement point P5 is the center of the rectangle. That is, measurement point P5 is located at the point where L2 = La / 2 and L3 = Lb / 2. Measurement points P6 and P7 are located a distance L1 inward from both outer edges of the side with length La and at the center of length Lb. Measurement point P8 is the midpoint between measurement points P5 and P7. That is, L4 = (L2 - L1) / 2. When measuring on both sides, measurement samples are taken from eight locations on each of the first surface 10a and the second surface 10b, and analyzed by ICP-OES to determine the nickel content. The samples taken are material from a circular area with a radius of 5 mm, for example, centered on the eight locations, to a depth of 3 mm. If the sample is too thin to take measurement samples to a depth of 3 mm from both sides, measurement samples may be taken by penetrating the thickness direction. The in-plane distribution of nickel content is evaluated by calculating the maximum, average, and standard deviation of the obtained numerical values. When the shape of the copper-molybdenum composite 10 as the sample is not rectangular, eight or more measurement points may be determined in accordance with the above-mentioned method for selecting measurement points.

[0053] <First Heat Sink> The heat sink 110 according to the embodiment of the present disclosure is a heat sink using the copper molybdenum composite 10 described above. The copper molybdenum composite 10 may be used as the heat sink 110 in its original form. Alternatively, the copper molybdenum composite 10 may be subjected to processes such as rolling or cutting to form the heat sink 110. The heat sink 110 can be used for heat dissipation purposes for semiconductor components and the like. FIGS. 1 and 2 schematically show the shape of the copper molybdenum composite 10 as well as the shape of the heat sink 110. That is, the copper molybdenum composite 10 and the heat sink 110 have a common appearance as plate-like bodies, except for the dimensions. The first surface 10a and the second surface 10b of the copper molybdenum composite correspond to the first surface 110a and the second surface 110b of the heat sink 110. The heat sink 110 typically has a thermal conductivity in the thickness direction at room temperature of 150 W / m·K or more and 320 W / m·K or less, and a linear expansion coefficient in the in-plane direction parallel to the surface from room temperature to 800°C of 6.0 ppm / K or more and 14.0 ppm / K or less.

[0054] Heat sinks can be used, for example, to prevent overheating of semiconductor elements. If there is a large difference in the linear expansion coefficient between the heat sink and the semiconductor element or an insulator such as a ceramic package, thermal stress occurs at the interface between them, which can lead to damage to the semiconductor element. For this reason, it is desirable for the linear expansion coefficient of the heat sink to be close to that of the semiconductor element or the insulator. A composite material made of copper, which has high thermal conductivity, and molybdenum, which has a low linear expansion coefficient, is suitable as a heat sink.

[0055] The thermal conductivity of copper is higher than that of molybdenum. The thermal conductivity of copper is, for example, 398 W / m·K at room temperature. "Room temperature" means 27°C. The linear expansion coefficient of molybdenum is, for example, 0.5 times or less that of copper. The linear expansion coefficient of molybdenum is, for example, 5.7 ppm / K. In this application, the "linear expansion coefficient" refers to the linear expansion coefficient in the temperature range from room temperature to 800°C.

[0056] (Method for Measuring Thermal Conductivity) The thermal conductivity of the heat sink 110 containing the copper-molybdenum composite 10 is calculated based on the thermal diffusion coefficient of the heat sink 110 and the volume ratio and specific heat of each constituent material of the heat sink 110. The thermal diffusion coefficient of the heat sink 110 is measured using a laser flash method. A NETZSCH LFA457 MicroFlash is used as the thermal diffusion coefficient measurement device. A sample used for measuring thermal conductivity is cut from the heat sink 110. The planar shape of the sample is a circle with a diameter of 10 mm. The specific heat of the material constituting the heat sink 110 is determined based on "Metal Data Book, Fourth Edition" (2004, Maruzen Publishing), edited by the Japan Institute of Metals. Prior to measuring the thermal conductivity of the heat sink 110, the thermal conductivity of pure copper of the same shape is measured under the same conditions, and the measurement results are used as a reference to correct the measurement results. Three samples of the same material are prepared. The average value of the thermal conductivities measured for the three samples was determined as the thermal conductivity of the heat sink 110. Here, the specific heat of copper is 386 J / (kg·K), and the specific heat of molybdenum is 251 J / (kg·K).

[0057] (Method for Measuring Linear Expansion Coefficient) The linear expansion coefficient of the heat sink 110 is calculated by measuring the expansion displacement in a direction perpendicular to the thickness direction of the heat sink 110 in a temperature range from room temperature to 800°C using a TD5000SA manufactured by Bruker AXS. A sample used to measure the linear expansion coefficient of the heat sink 110 is cut out from the heat sink 110. The planar shape of the sample is a rectangle measuring 5 mm x 15 mm. Three samples made of the same material are prepared. The average value of the linear expansion coefficients measured for the three samples is taken as the linear expansion coefficient of the heat sink 110.

[0058] <<Second Heat Sink>> A heat sink 120 according to another embodiment of the present disclosure is a laminate using the above-described copper molybdenum composite 10. Figures 4A and 4B are diagrams each showing an example of a cross section perpendicular to the surface of the heat sink 120. The heat sink 120 is a laminate in which at least one first material layer 121 made of the copper molybdenum composite 10 and a second material layer 122 containing copper as a main component are alternately stacked in the thickness direction.

[0059] (Number of Layers) The total number of layers of the first material layer 121 and the second material layer 122 is not limited. FIG. 4A illustrates an example of a laminate with a total of three layers. The heat sink 120 in FIG. 4A is a laminate with the second material layer 122, the first material layer 121, and the second material layer 122 stacked in that order from the surface. FIG. 4B illustrates a laminate with a total of five layers as a diagram showing another example of the cross section of the heat sink 120. The heat sink 120 in FIG. 4B is a laminate with the second material layer 122, the first material layer 121, the second material layer 122, the first material layer 121, and the second material layer 122 stacked in that order from the surface.

[0060] The layers constituting the first surface 120a and the second surface 120b may be made of the same material. In this configuration, the number of layers is odd, resulting in a stacked body that is symmetrical from the center in the stacking direction. This symmetrical configuration in the stacking direction makes it less likely to warp or deform due to temperature changes.

[0061] The first material layer 121 is a layer made of the copper-molybdenum composite 10 described above. The first material layer 121 has a lower linear expansion coefficient than the second material layer 122. The second material layer 122 is a layer made of a material whose main component is copper. The second material layer 122 has a higher thermal conductivity than the first material layer 121. The heat sink 120 has a layered structure in which the first material layers 121 and the second material layers 122 are alternately stacked. Adjacent first material layers 121 and second material layers 122 are in contact with each other. This layered structure allows the heat sink 120 to achieve both high thermal conductivity and a low linear expansion coefficient.

[0062] The thermal conductivity of the second material layer 122 may be, for example, 300 W / m·K or more, or even 350 W / m·K or more. The method for measuring thermal conductivity is as described above. The second material layer 122 is a metal material primarily composed of copper. A "metal material primarily composed of copper" refers to a metal material having a copper content of 50% by mass or more. In other words, when the total mass of the second material layer 122 is 100% by mass, the copper content in the second material layer 122 is 50% by mass or more. The copper content may also be 70% by mass or more. The second material layer 122 is made of, for example, pure copper. "Pure copper" refers to a metal material composed of copper and unavoidable impurities constituting the remainder. Pure copper has a copper content of 99% by mass or more, particularly 99.9% by mass or more.

[0063] The thickness of each layer will be described with reference to FIG. 4A . In FIG. 4A , the sum of the thickness T1 of the first material layer 121 and the thicknesses T21 and T22 of the second material layer 122 is the thickness T0 of the heat sink 120. The thicknesses T21 and T22 of the second material layer 122 may be, for example, 0.025 mm or more. When the thicknesses T21 and T22 are 0.025 mm or more, the thermal conductivity of the heat sink 120 is likely to be high. The thicknesses T21 and T22 may each be 40% or less of the thickness T0. When the thicknesses are 40% or less, the thickness of the first material layer 121 does not become too small. This allows the linear expansion coefficient of the heat sink 120 to be low. The lower limits of the thicknesses T21 and T22 may be 0.035 mm or more, or 0.045 mm or more. The upper limits of the thicknesses T21 and T22 may be less than 40% or less than 30% of the thickness T0. The above upper and lower limits can be combined in any manner.

[0064] When the heat sink 120 includes multiple second material layers 122, all of the second material layers 122 may have the same thickness, or some of the second material layers 122 may have different thicknesses. In FIG. 4B , thickness T21 and thickness T22 may be the same or different. Here, if thickness T21 and thickness T22 are the same, a stacked body symmetrical in the stacking direction is obtained. Having a symmetrical configuration in the stacking direction makes it less likely to warp or deform with temperature changes. In the example shown in FIG. 4B , of the three second material layers 122, the second material layer 122 located at the center of the heat sink 120 in the thickness direction is depicted with a thickness T23 that is thinner than the thicknesses T21 and T22. By varying the thicknesses in this way, it is easy to adjust the thermal conductivity of the heat sink 120.

[0065] The thickness T1 of the first material layer 121 may be, for example, 0.05 mm or more. When the thickness T1 is 0.05 mm or more, the first material layer 121, which has a relatively low linear expansion coefficient, has an appropriate thickness. Therefore, the linear expansion coefficient of the heat sink 120 is likely to be low. The thickness T1 may be 50% or less of the thickness T0 of the heat sink 120. When the thickness T1 is 50% or less of the thickness T0, the second material layer 122 can have an appropriate thickness. Therefore, the thermal conductivity of the heat sink 120 is likely to be high. The lower limit of the thickness T1 may be 0.10 mm or more, or may be 0.15 mm or more. The upper limit of the thickness T1 may be 40% or less of the thickness T0, or may be 30% or less of the thickness T0. The above upper and lower limits can be combined arbitrarily.

[0066] Regarding the thicknesses of the multiple first material layers 121, the thickness T1 of all the first material layers 121 may be the same, or some may differ. In the example shown in FIG. 4B , the thickness T1 of each of the two first material layers 121 is the same. When the thickness T1 of the multiple first material layers 121 is the same, a stacked body that is symmetrical in the stacking direction is obtained. The symmetrical configuration in the stacking direction makes it less likely to warp or deform due to temperature changes.

[0067] In the above description, the thickness can be measured by a known measurement method such as a contact method or a non-contact method. In the embodiments of the present disclosure, the thickness is the average value of values ​​obtained by measuring three or more points on the same sample.

[0068] A heat sink composed of multiple layers can easily achieve the desired thermal conductivity and linear expansion coefficient by changing the thickness and number of each layer. Therefore, a heat sink having a stable, high thermal conductivity and a linear expansion coefficient close to that of semiconductors and ceramic insulators can be obtained. The heat sink 120 according to the embodiment may have a thermal conductivity in the thickness direction at room temperature of 180 W / m·K or more and 350 W / m·K or less. The linear expansion coefficient from room temperature to 800°C in an in-plane direction perpendicular to the thickness direction may be 6.5 ppm / K or more and 14.0 ppm / K or less.

[0069] In the heat sink 120, which is a laminate, the layers constituting the first surface 120a and the second surface 120b, which are the end faces in the thickness direction, may both be the second material layer 122. The end faces in the thickness direction of the heat sink 120 become surfaces that transfer heat by coming into contact with other members. By using the second material layer 122, which has a high thermal conductivity, as the material constituting the end faces, it becomes easier to absorb heat from other members that it comes into contact with, or to radiate heat to other members.

[0070] <<Modified Heat Sink>> A heat sink 130 according to another embodiment of the present disclosure is a heat sink having a metal coating on at least a portion of the first surface, the second surface, or both surfaces of the heat sink 110 or the heat sink 120 described above. FIG. 5 is a cross-sectional schematic diagram showing an example of the heat sink 130. The heat sink 130 of FIG. 5 has a metal coating 50 on the entire first and second surfaces. The metal coating 50 may be made of at least one metal selected from the group consisting of nickel, nickel-phosphorus, copper, silver, gold, and platinum, or an alloy thereof. The material of the metal coating 50 is a metal that has good bonding properties with a bonding material used to bond the heat sink 130 to a heat dissipation target. The bonding material may be a brazing material or a solder material, as described below. Molybdenum has poorer bonding properties with bonding materials than copper and the like. Therefore, providing a metal coating 50 is effective for heat sinks with a high molybdenum content on the end surfaces.

[0071] Semiconductor Device A semiconductor device 300 according to an embodiment of the present disclosure will be described with reference to Figures 6A and 6B. Figure 6A is an exploded perspective view of the semiconductor device 300. Figure 6B is a schematic cross-sectional view of the semiconductor device of Figure 6A. The semiconductor device 300 includes a heat sink 310, a semiconductor element 320, a case member 330, a lid 350, and terminals 360a and 360b.

[0072] The heat sink 310 constitutes a heat spreader of the semiconductor device 300. The heat sink 310 is the heat sink 110, the heat sink 120, or the heat sink 130 described above. The heat sink 310 includes the copper-molybdenum composite 10. The semiconductor element 320 is a heat source during operation. The semiconductor element 320 is disposed on a first surface 310a of the heat sink.

[0073] A heat dissipation member 340 is attached to the second surface 310b of the heat dissipation plate 310. The heat dissipation member 340 is, for example, a metal plate having a flow path formed therein through which a refrigerant flows. The heat dissipation member 340 is not limited to this. The heat dissipation member 340 may be, for example, a cooling fin. A heat transfer member (not shown) may be interposed between the heat dissipation member 340 and the second surface 310b.

[0074] The case member 330 is made of an insulating material such as a ceramic material. The ceramic material is, for example, alumina (Al ２ O ３ ) The case member 330 in this example has a rectangular frame shape. The case member 330 is disposed on the first surface 310a so as to surround the semiconductor element 320. The lid 350 is made of, for example, a ceramic material or a metal material. The lid 350 closes the opening of the case member 330.

[0075] Terminals 360a and 360b are inserted into case member 330. First ends of terminals 360a and 360b are located within a space defined by first surface 310a, case member 330, and lid 350. Second ends of terminals 360a and 360b are located outside the space. Terminals 360a and 360b are made of, for example, a metal material. The metal material is, for example, Kovar. Although not shown, the first ends of terminals 360a and 360b are electrically connected to semiconductor element 320. Semiconductor device 300 is electrically connected to a device or circuit other than semiconductor device 300 via the second ends of terminals 360a and 360b.

[0076] Referring to FIG. 6B , the semiconductor element 320 and the heat sink 310 are bonded together by a bonding material 370a. The case member 330, which is an insulator, and the heat sink 310 are bonded together by a bonding material 370b. The bonding materials 370a and 370b are materials containing lead-containing solder, lead-free solder, brazing material, silver sinter, or copper sinter. The bonding materials 370a and 370b may be the same material or different materials. When the case member 330 and the heat sink 310 are bonded together by the bonding material 370b, the heat sink 310 is exposed to high temperatures, for example, approximately 800°C. The difference between the linear expansion coefficient of the heat sink 310 containing the copper-molybdenum composite 10 and the linear expansion coefficient of the alumina constituting the case member 330 is small. Therefore, thermal stress generated at the interface between the heat sink 310 and the case member 330 is small, making damage to the semiconductor element 320 and other components less likely.

[0077] <Manufacturing Method of Copper-Molybdenum Composite> A manufacturing method of a copper-molybdenum composite according to this embodiment will be described with reference to FIG. 9 . The manufacturing method of a copper-molybdenum composite includes step A (step S101) of preparing a composite material, step B (step S102) of rolling the composite material, and step C (step S103) of heat-treating the composite material. Step A is a step of infiltrating copper into a molybdenum compact obtained by compression-molding molybdenum powder (step S101a). Alternatively, step A may be a step of infiltrating copper into a molybdenum compact obtained by compression-molding a mixed powder of copper and molybdenum (step S101b). In step B, the composite material is warm-rolled to form a rolled plate. In step C, the rolled plate is passed through a heat treatment furnace while being moved at a feed rate of 100 mm / min or less. The heat treatment furnace is in a reducing atmosphere. The composite material is held in the heat treatment furnace at a temperature of 700°C to 1000°C for 0.3 hours or more.

[0078] (Step A) Step A comprises the steps of preparing a molybdenum compact and infiltrating the molybdenum compact with copper. The first method S101a for preparing the molybdenum compact involves pressing molybdenum powder having an average particle size of about 2 μm to 6 μm in a hydraulic or isostatic press (CIP) at 5 ton / cm. ２ The second method S101b for preparing a molybdenum powder compact is a method of mixing molybdenum powder with an average particle size of about 2 μm to 6 μm and 5 mass % or more of copper powder, and pressing the mixture in a hydraulic or isostatic press at a pressure of 5 ton / cm. ２ This method involves forming a plate-shaped compact using the following pressure.

[0079] The above-mentioned powder compact is a plate-like body having a first surface and a second surface, which are opposite each other. A copper plate is placed on either the first surface or the second surface, and the body is placed in a furnace with a reducing or inert atmosphere. The temperature inside the furnace is heated to a temperature of 1083°C or higher and 1300°C or lower, thereby melting the copper. The molten copper is impregnated (hereinafter sometimes referred to as "infiltration") into the molybdenum powder compact. The copper infiltrated into the powder compact forms a copper-molybdenum composite as a composite material. Copper can be infiltrated at 20% by mass or higher and 80% by mass or lower based on the total mass of the copper-molybdenum composite. The second method S101b described above makes it easier to produce a copper-molybdenum composite with a higher copper ratio than the first method S101a.

[0080] The molybdenum powder used as raw material has an FSSS particle size range of 2 μm to 6 μm according to the Fischer method. If the particle size is too fine, the powder volume increases, and a high compaction pressure is required to achieve the desired apparent density. If the particle size is too coarse, compactibility decreases and the periphery of the green compact becomes prone to collapse.

[0081] The infiltration temperature is in the range of 1083° C. to 1300° C. The temperature may be 1150° C. or higher. If the temperature is in this range, a copper-molybdenum composite can be stably obtained without generating voids.

[0082] (Step B) In step B, the composite material infiltrated with copper in step A is heated to a temperature of 150°C to 300°C and warm-rolled into a rolled plate. Typically, the composite material produced in step A has a thickness of about 10 mm to 20 mm. The thickness of the rolled plate is about 1 mm to 5 mm. The thicknesses of the respective components are not limited to the above values.

[0083] (Step C) Step C is a step of subjecting the composite material, which is the copper-molybdenum composite obtained in Step A, to a heat treatment. The heat treatment of the present disclosure will be explained using Figures 7 and 8. Figures 7 and 8 schematically show the positional relationship between a heating furnace and the composite material to be heated.

[0084] 7 is a diagram illustrating heat treatment in a batch furnace, which is a conventional heat treatment method. A composite material 150 is placed in an internal space 251 of a heating furnace 250. The internal space 251 is a reducing atmosphere. The internal space 251 is maintained at a constant temperature of 700°C or higher and 1000°C or lower. After being introduced into the internal space 251 at the above temperature, the composite material 150 is maintained for a predetermined time.

[0085] 8 is a diagram illustrating heat treatment in a belt furnace, which is a heat treatment method according to the present disclosure. The heating furnace 210 has an internal space 211. The heating furnace 210 has openings 212 and 213. The internal space 211 is a reducing atmosphere. Most of the internal space 211, except for the vicinity of the openings 212 and 213, is maintained at a constant temperature of 700° C. or higher and 1000° C. or lower.

[0086] A belt 220 is provided to transport the heat treatment target through the openings 212 and 213. The composite material 150 to be heat treated is placed on the belt 220. The composite material 150 is transported through the heating furnace 210, with the opening 212 as the entrance and the opening 213 as the exit. The thick arrow in the figure indicates the direction of transport, and V indicates the transport speed. The transport speed V is 100 mm / min or less. That is, the composite material 150 enters the internal space 211 through the opening 212 at a speed of 100 mm / min or less, passes through the internal space 211, and is then discharged from the furnace through the opening 213. During the transport process, the composite material 150 is held in the internal space 211 for a certain period of time. That is, the composite material 150 is held at a temperature of 700°C or higher and 1000°C or lower for 0.3 hours or longer. There is no particular upper limit to the holding time, but it may be 3 hours or shorter, or 1 hour or shorter, taking into account factors such as production efficiency.

[0087] The inventors discovered that nickel enrichment occurs near the center of the composite material produced by process A as the infiltrated copper cools and solidifies from the periphery, and conceived that this state could be changed by adjusting the heat treatment conditions. It is believed that the enriched nickel is diffused and evenly dispersed within the surface by heat treatment in a belt furnace at a speed of 100 mm / min or less. On the other hand, in heat treatment using a belt furnace with a high feed rate or a batch furnace, the enriched nickel remains near the center of the composite material 150. Under the above-described heat treatment conditions using a belt furnace, the composite material 150 is heated sequentially from one end as it is transported into the internal space 211 through the opening 212. In conventional heat treatment using a batch furnace, the entire composite material 150 is heated to a constant temperature, so the enriched nickel does not disappear. On the other hand, in the heat treatment disclosed herein, the treatment proceeds at a slower rate from the end, resulting in a dispersed nickel enrichment in the central portion. In this way, a copper molybdenum composite 10 can be obtained in which the standard deviation of the in-plane distribution of the nickel content is 5 ppm or less.

[0088] <<First Method for Manufacturing Heat Sink>> The copper molybdenum composite 10 produced by the above-described method for manufacturing a copper molybdenum composite can be used as a heat sink 110 in its original form. Furthermore, processing such as rolling, cutting, or punching may be performed to obtain a heat sink 110 of a desired size. FIG. 10 is a flow diagram showing a method for manufacturing a heat sink 110. The method for manufacturing a heat sink 110 includes a step (S201) of preparing the copper molybdenum composite 10 obtained by the above-described method for manufacturing a copper molybdenum composite, a step (S202) of rolling the copper molybdenum composite 10, and a step (S203) of cutting the rolled copper molybdenum composite 10. Of the above steps, the rolling and cutting steps are performed as needed.

[0089] The copper-molybdenum composite 10 is, for example, a plate-like body with a long side of 100 mm or more, a short side of 100 mm or more, and a thickness of 2 mm or more. By rolling this plate-like body, a heat sink 110 of a desired thickness, for example, a thickness of 1 mm or less, can be obtained. The rolling processing rate can be 20% or more and 95% or less. Increasing the processing rate increases the length of the molybdenum structure, which has the effect of reducing the linear expansion coefficient. On the other hand, if the processing rate is too high, cracks are likely to occur due to work hardening.

[0090] Heat sinks using the copper-molybdenum composite 10 of the present disclosure do not exhibit nickel enrichment and have uniform in-plane thermal conductivity. Therefore, even when used as large-area heat sinks, they do not contain areas with high thermal conductivity locally, making them less likely to develop heat spots. Furthermore, even when cut into multiple small heat sinks, each heat sink has uniform thermal conductivity. In other words, there is little variation in the quality of the multiple heat sinks.

[0091] Such heat sink 110 can have a thermal conductivity in the thickness direction at room temperature of 150 W / m·K or more and 320 W / m·K or less. Also, heat sink 110 can have a linear expansion coefficient from room temperature to 800°C in an in-plane direction parallel to the first surface of 6.0 ppm / K or more and 14.0 ppm / K or less.

[0092] <<Second Method for Manufacturing Heat Sink>> The copper molybdenum composite 10 produced by the above-described method for manufacturing a copper molybdenum composite can be laminated with another plate-shaped body to form a heat sink 120. The following description will be made again with reference to FIG. 4A or FIG. 4B . The laminated heat sink 120 may be processed, such as by rolling, cutting, or punching, to form the heat sink 110 of a desired size. FIG. 11 is a flow diagram showing a method for manufacturing the heat sink 120. The method for manufacturing the heat sink 120 includes steps of preparing the copper molybdenum composite 10 that will become the first material layer 121 and a plate-shaped body that will become the second material layer 122 (S301 and S302), a step of laminating the first material layer 121 and the second material layer 122 to obtain a laminate (S303), a step of rolling the laminate (S304), and a step of cutting the rolled laminate (S305). Of the above steps, the rolling step and the cutting step are steps that are carried out as necessary.

[0093] In the preparation step, at least one copper-molybdenum composite 10 is prepared as a first material layer 121. A plurality of second material layers 122 are also prepared. The second material layers 122 are made of a metal material containing copper as a main component. The thickness and size of each of the plate-like bodies that will become the first material layer 121 and the second material layer 122 are selected appropriately depending on the application of the heat sink 120 and the heat dissipation characteristics that are desired to be exhibited.

[0094] The lamination process is a hot-press process in which multiple prepared plate-shaped bodies are stacked and heated and pressurized. The first material layer 121 and the second material layer 122 are alternately stacked. For example, if the layers that form the first surface 120a and the second surface 120b are the second material layer 122, copper plates, which are the second material layer 122, are placed at the bottom and top of the stack, respectively. The lamination process is performed, for example, inside a mold made of graphite. The stacked body arranged inside the mold is heated to a temperature below the melting point of the second material layer 122 and at which the first material layer 121 is sufficiently softened. The heating temperature is, for example, 1000°C. Pressurization is performed in the direction along the thickness of the stack. The pressure is, for example, 50 MPa or more. The first material layer 121 and the second material layer 122 are bonded to each other by the above-mentioned heating and pressing. The processes of rolling and cutting the bonded stack are the same as those in the above-mentioned first heat sink manufacturing method.

[0095] 5 , the above-described metal coating 50 may be formed on the first surface, the second surface, or both of the heat sink 110 or the heat sink 120 obtained by the above-described manufacturing method. The metal coating 50 may be formed by a known method such as electrolytic nickel plating using a Watts bath.

[0096] Test Example 1 A heat sink was prepared as follows and its characteristics were evaluated.

[0097] (Production of Composite Material 1) Molybdenum powder with an FSSS particle size of 3.9 μm was prepared by the Fischer method. This molybdenum powder contained 4 ppm of nickel as an inevitable impurity. Using a press, the molybdenum powder was molded at a pressure of 200 MPa into a size of 171 mm in length, 143 mm in width, and 13 mm in thickness. This molded body was placed in a sintering furnace in a hydrogen atmosphere and sintered at a sintering temperature of approximately 1150°C. The densities of the obtained sintered bodies ranged from 5.00 to 8.00 g / cm. ３ The density of the sintered body was adjusted by changing the sintering temperature.

[0098] A copper plate with the same surface size was placed on one side of the obtained sintered body. The copper plate was made of copper with a purity of 99.9% or more. The sintered body with the copper plate placed on it was placed in a sintering furnace in a hydrogen atmosphere and held at a temperature of approximately 1200°C for 2.5 hours, thereby infiltrating copper into the pores of the molybdenum porous body. After that, excess copper remaining on the surface was removed by polishing, and a composite material was obtained.

[0099] By preparing multiple molybdenum porous bodies with different sintered densities, multiple composite materials with copper contents ranging from 20% by mass to 50% by mass were obtained. In the following description, the material produced by the above method will be referred to as PCM, followed by a number indicating the copper content. For example, "PCM20" indicates a copper-molybdenum composite obtained by the above manufacturing method, with a copper content of approximately 20% by mass relative to the total mass of the copper-molybdenum composite. In this example, five types of samples were prepared: PCM20, PCM30, PCM35, PCM40, and PCM50.

[0100] (Production of Composite Material 2) Molybdenum powder with an FSSS particle size of 3.9 μm obtained by the Fischer method and copper powder with an FSSS particle size of 2 μm were prepared. The prepared molybdenum powder and copper powder were mixed so that each was 50 mass%. This mixed powder contained 2 ppm of nickel as an unavoidable impurity. Using a press, the molybdenum powder was molded at a pressure of 200 MPa into a size of 171 mm in length, 143 mm in width, and 13 mm in thickness. The molded body was placed in a sintering furnace in a hydrogen atmosphere and sintered at a sintering temperature of approximately 1150°C. The densities of the obtained sintered bodies ranged from 6.00 to 9.40 g / cm. ３ The density of the sintered body was adjusted by changing the sintering temperature.

[0101] The resulting sintered body was infiltrated with copper using the same method as in Composite Material Production 1. By preparing multiple porous molybdenum bodies with different sintered body densities, multiple composite materials with copper contents ranging from 50% to 80% by mass were obtained. In the following description, the material produced by the above method will be referred to as RCM, followed by a number indicating the copper content. For example, "RCM50" indicates a copper-molybdenum composite obtained by the above method, in which the copper content relative to the total mass of the copper-molybdenum composite is approximately 50% by mass. In this example, four types of samples, RCM50, RCM60, RCM70, and RCM80, were prepared.

[0102] (Rolling) Each of the composite materials obtained above was warm-rolled at a temperature of 160° C. to a thickness of 2.0 mm to obtain a rolled plate.

[0103] (Heat Treatment 1 in a Belt Furnace) The above-mentioned rolled plate was placed on the belt of a belt furnace. The temperature inside the belt furnace was approximately 700°C. The composite material was loaded into the belt furnace by moving the belt at a feed rate of 100 mm / min. The loaded composite material moved through the belt furnace at a constant speed. The composite material was held in the atmosphere inside the belt furnace for approximately 0.3 hours and then removed from the belt furnace to obtain a heat sink made of a copper-molybdenum composite. The obtained heat sinks are shown in Table 1. There were nine types of heat sinks, Sample No. 11 to Sample No. 19. Each sample was the above-mentioned PCM or RCM with a different copper content.

[0104] (Heat Treatment 2 in a Belt Furnace) The above-mentioned rolled plate was placed on the belt of a belt furnace, and the belt was moved at a feed rate of 150 mm / min to obtain a heat sink made of a copper-molybdenum composite. The conditions in the belt furnace other than the feed rate were the same as those in Heat Treatment 1 in a Belt Furnace described above. The obtained heat sinks are shown in Table 2. There were nine types of heat sinks, Sample No. 101 to Sample No. 109. Each sample was the above-mentioned PCM or RCM, but with a different copper content.

[0105] (Heat Treatment in a Batch Furnace) The above-mentioned rolled plate was placed in a batch furnace. The batch furnace was heated to approximately 700°C. The composite material was held in the batch furnace for approximately 0.3 hours to obtain a heat sink made of a copper-molybdenum composite. The obtained heat sinks are shown in Table 3. There were nine types of heat sinks, Sample No. 111 to Sample No. 119. Each sample was the above-mentioned PCM or RCM, but with a different copper content.

[0106]

[0107]

[0108]

[0109] (Evaluation) From each of the prepared specimens, samples required for measurement were obtained, and the nickel content, thermal conductivity, and linear expansion coefficient were measured.

[0110] (In-plane Distribution of Nickel Content) The in-plane distribution of nickel content was measured based on the above-described measurement method. Tables 1 to 3 show the nickel content (referred to as "Ni content" in the tables) measured at the above-described eight locations (P1 to P8) for each sample, as well as the maximum value and standard deviation of the eight values. Referring to FIG. 3 , the sample size and measurement locations for the nickel content measurement method are as follows: La = 930 mm, Lb = 170 mm, L1a = 33 mm, L1b = 5 mm, L2 = 465 mm, L3 = 85 mm, and L4 = 248 mm. Furthermore, a 3 mm diameter drill was used to drill holes at each measurement location P1 to P8, and the resulting cuttings were used as samples. Note that the notation "<3" for the Ni content indicates that the Ni content is less than 3 ppm, which is the measurement limit. Data less than 3 ppm was used in statistical calculations as 2 ppm.

[0111] The data shown in Tables 1, 2, and 3 include evaluation results for copper-molybdenum composites with the same copper content but different heat treatment conditions. Differences can be understood by comparing the corresponding samples. For Samples No. 11 to No. 19, which underwent Heat Treatment 1 in a belt furnace, the Ni content at measurement points P1 to P8 was nearly identical. That is, the Ni amount was similar at every location. This uniformity of Ni content within the plane is reflected by the small standard deviation on each plane. On the other hand, for Samples No. 101 to No. 109, which underwent Heat Treatment 2 in a belt furnace, the Ni content at P5, the center of the plane, was higher among measurement points P1 to P8. This indicates that nickel enrichment occurred in the center of the plane. Furthermore, for Samples No. 111 to No. 119, which underwent Heat Treatment in a batch furnace, the Ni content at P5, the center of the plane, was higher among measurement points P1 to P8. In other words, it can be seen that nickel is concentrated in the center of the surface. From the above, according to the heat treatment conditions of the present disclosure using a belt furnace, it is possible to obtain a heat sink plate with a uniform nickel content in the surface.

[0112] (Thermal Conductivity) The thermal conductivity of each sample was measured based on the measurement method described above. For each sample, the thermal conductivity was measured on samples cut out by wire electrical discharge machining from three locations within the surface. One of the three locations was near the center of the surface, which was measurement location P5 in the measurement of the nickel content. The minimum value of the thermal conductivity measured at the three locations was taken as the thermal conductivity of that sample. The thermal conductivity of each sample in Tables 1, 2, and 3 was within the range required for a heat sink. When comparing the samples in Tables 2 and 3 with samples having the same copper content as Table 1, the samples in Tables 2 and 3 had lower values ​​than the sample in Table 1. This is thought to be because the samples shown in Tables 2 and 3 have a high nickel content near the center of the surface, which is where the thermal conductivity is low.

[0113] (Linear expansion coefficient) The linear expansion coefficient of each sample was measured based on the measurement method described above. For each sample, the linear expansion coefficient was measured for samples cut out by wire electric discharge machining from three locations within the surface. One of the three locations was near the center of the surface, which was measurement location P5 in the measurement of the nickel content. No significant difference was observed between the linear expansion coefficients of the three measured locations. Therefore, the average value of the measurements from the three locations was used as the linear expansion coefficient of that sample. The linear expansion coefficients of each sample in Tables 1, 2, and 3 were all within the range required for a heat sink. Even when comparing each sample in Tables 2 and 3 with samples having the same copper content as in Table 1, no effect on the linear expansion coefficient due to differences in heat treatment conditions was confirmed.

[0114] Test Example 2 A heat sink as a laminate was fabricated as follows, and its properties were evaluated. The heat sink used in the test was a three-layer laminate. The laminate had a structure in which one layer of a copper-molybdenum composite was sandwiched between two copper layers constituting a first surface and a second surface.

[0115] (Production of copper-molybdenum composites) Composite materials PCM20, PCM30, PCM35, PCM40, PCM50, RCM50, RCM60, RCM70, and RCM80 were prepared by the method of Test Example 1.

[0116] (Rolling) Each of the composite materials obtained above was warm-rolled at a temperature of 160° C. to a thickness of 2.0 mm to obtain a rolled plate of a copper-molybdenum composite.

[0117] (Manufacture of Laminate) For each of the copper-molybdenum composites described above, a copper plate was laminated on each of the first and second surfaces, which were both end faces. The copper plate material used was copper with a purity of 99.9% or higher. The copper plate thickness was 0.5 mm. The laminated bodies were hot-rolled at a temperature of 840°C in a hydrogen atmosphere to be roll-bonded to each other. Next, copper oxides formed on the surfaces were removed with a brush or the like, and the bodies were annealed at 850°C for 0.3 hours. The annealed bonded bodies were cold-rolled to a total thickness of 2 mm to obtain a heat sink.

[0118] The obtained heat sink is a laminate of, from the surface layer, a copper layer, a copper molybdenum composite layer, and a copper layer, and will be referred to as "CPC" in the following description. The thickness ratio of the copper layer to the copper molybdenum composite layer is expressed numerically. For example, "CPC141" indicates the above three-layer structure, in which the three layers are stacked in the order of thickness ratio 1:4:1. Furthermore, in the following description, "CPC141(50)" indicates a laminate with the above thickness ratio, in which the copper molybdenum composite layer is PCM50 or RCM50. CPC is a registered trademark of A.L.M.T. Corporation, which sells heat sinks with a copper-copper molybdenum-copper laminate structure.

[0119] (Heat Treatment 1 in a Belt Furnace) Heat treatment in a belt furnace was performed in the same manner as in Test Example 1 to obtain a heat sink, which was a laminate using a copper-molybdenum composite. The heat treatment conditions were a temperature of approximately 700°C, a feed rate of 100 mm / min, and a holding time of approximately 0.3 hours. The obtained heat sinks are shown in Table 4. There were nine types of heat sinks, Sample No. 21 to Sample No. 29. Each sample had the same laminate structure, but the copper content of the copper-molybdenum composite layer was different.

[0120] (Heat treatment 2 in a belt furnace) Heat treatment was performed in a belt furnace under the conditions of a temperature of approximately 700°C, a feed rate of 150 mm / min, and a holding time of approximately 0.3 hours to obtain heat sinks. The obtained heat sinks are shown in Table 5. There were nine types of heat sinks, Sample No. 201 to Sample No. 209. Each sample had the same laminate structure, but the copper content of the copper-molybdenum composite layer was different.

[0121] (Heat Treatment in a Batch Furnace) Heat treatment in a batch furnace was performed under the same conditions as in Test Example 1 to obtain heat sinks that were laminates using copper-molybdenum composites. The obtained heat sinks are shown in Table 6. There were nine types of heat sinks, Sample No. 211 to Sample No. 219. Each sample had the same laminate structure but differed in the copper content of the copper-molybdenum composite layer.

[0122] (Evaluation) Evaluation was carried out using the same methods and for the same evaluation items as in Test Example 1. The measurement results are shown in Tables 4, 5 and 6.

[0123]

[0124]

[0125]

[0126] (In-plane distribution of nickel content) For each sample, the nickel content measured at the above-mentioned eight locations (P1 to P8), as well as the maximum value and standard deviation of the eight values, are shown in Tables 4, 5, and 6. The size of the test piece, the measurement locations, and the method of taking the samples were the same as in Test Example 1.

[0127] The data shown in Tables 4, 5, and 6 show the evaluation results of samples with copper-molybdenum composite layers having the same copper content but different heat treatment conditions. Comparing the corresponding samples reveals the following differences. For Samples No. 21 to No. 29, which underwent Heat Treatment 1 using a belt furnace, the Ni content at measurement points P1 to P8 was nearly identical. That is, the Ni content was similar at every location. This uniformity of Ni content across the surface is reflected by the small standard deviation across each surface. On the other hand, for Samples No. 201 to No. 209, which underwent Heat Treatment 2 using a belt furnace, the Ni content at P5, the center of the surface, was greater than that at measurement points P1 to P8. This indicates that nickel enrichment occurred in the center of the surface. Furthermore, for Samples No. 211 to No. 219, which underwent Heat Treatment 1 using a batch furnace, the Ni content at P5, the center of the surface, was greater than that at measurement points P1 to P8. In other words, it can be seen that nickel is concentrated in the center of the surface. From the above, according to the heat treatment conditions of the present disclosure using a belt furnace, it is possible to obtain a heat sink plate with a uniform nickel content in the surface.

[0128] (Thermal Conductivity) As in Test Example 1, the minimum value of the thermal conductivity measured at three locations for each sample was used as the thermal conductivity of that sample. The thermal conductivity of each sample in Tables 4, 5, and 6 was within the range required for a heat sink. When comparing the samples in Tables 5 and 6 with corresponding samples with the same copper content as Table 4, the samples in Tables 5 and 6 had lower values ​​than the sample in Table 4. This is thought to be due to the fact that the samples shown in Tables 5 and 6 have a high nickel content near the center of the surface, resulting in low thermal conductivity in that area. The thermal conductivity of a laminated heat sink can be increased compared to that of a copper-molybdenum composite alone. The composite heat sink has a high copper content due to the copper layer on the surface. It can be seen that the heat sinks shown in Tables 4 to 6 have higher thermal conductivity than the heat sinks shown in Tables 1 to 3.

[0129] (Linear expansion coefficient) The linear expansion coefficient of each sample was measured in the same manner as in Test Example 1. The linear expansion coefficient of each sample in Tables 4, 5, and 6 was within the range required for a heat sink. Even when each sample in Tables 5 and 6 was compared with the corresponding sample having the same copper content as in Table 4, no effect of differences in heat treatment conditions on the linear expansion coefficient was confirmed.

[0130] It should be noted that the embodiments disclosed herein are illustrative in all respects and should not be considered restrictive. The present invention is not limited to these examples, but is defined by the scope of the claims, and is intended to include all modifications within the scope and meaning equivalent to the claims.

[0131] 10 Copper molybdenum composite 10a, 110a, 120a, 310a First surface 10b, 110b, 120b, 310b Second surface 50 Metal coating 110, 120, 130, 310 Heat sink 121 First material layer 122 Second material layer 150 Composite material 210, 250 Heating furnace 211, 251 Internal space 212, 213 Opening 220 Belt 300 Semiconductor device 320 Semiconductor element 330 Case member 340 Heat dissipation member 350 Lid 360a, 360b Terminal 370a, 370b Bonding material

Claims

1. A copper-molybdenum composite, comprising a plate-shaped body having a first surface and a second surface, a copper content of 20% by mass or more and 80% by mass or less, and a standard deviation of the in-plane distribution of the nickel content measured at multiple locations on the first surface and the second surface is 5 ppm or less.

2. The copper molybdenum composite according to claim 1, wherein the maximum nickel content is 20 ppm or less.

3. A heat sink made from the copper-molybdenum composite of claim 1 or 2, having a thermal conductivity in the thickness direction at room temperature of 150 W / m.K or more and 320 W / m.K or less, and a linear expansion coefficient from room temperature to 800°C in an in-plane direction perpendicular to the thickness direction of 6.0 ppm / K or more and 14.0 ppm / K or less.

4. The heat sink according to claim 3, wherein at least a portion of the first surface or the second surface is provided with a metal coating, the metal coating being made of at least one metal selected from the group consisting of nickel, nickel-phosphorus, copper, silver, gold, and platinum, or an alloy thereof.

5. A heat sink having a laminated structure of at least one first material layer made of the copper-molybdenum composite according to claim 1 or 2 and a second material layer whose main component is copper.

6. The heat sink according to claim 5, wherein the thermal conductivity in the thickness direction at room temperature is 180 W / m·K or more and 350 W / m·K or less, and the linear expansion coefficient from room temperature to 800°C in an in-plane direction perpendicular to the thickness direction is 6.5 ppm / K or more and 14.0 ppm / K or less.

7. A heat sink according to claim 5 or claim 6, wherein the layers constituting the upper and lower surfaces, which are the end surfaces in the thickness direction, are both layers of the second material.

8. A heat sink according to any one of claims 5 to 7, wherein at least a portion of the upper or lower surface, which is an end surface in the thickness direction, is provided with a metal coating, the metal coating being made of at least one metal selected from the group consisting of nickel, nickel-phosphorus, copper, silver, gold, and platinum, or an alloy thereof.

9. A semiconductor device comprising a heat sink according to any one of claims 3 to 8 and a semiconductor element, and a bonding material that bonds the semiconductor element and the heat sink, or between an insulator and the heat sink, wherein the bonding material includes any one of leaded solder, lead-free solder, brazing material, silver sintered material, and copper sintered material.

10. A method for producing a copper-molybdenum composite, comprising: a step of producing a composite material by either a method of infiltrating copper into a molybdenum powder compact, or a method of infiltrating copper into a powder compact consisting of a mixture of copper and molybdenum; and a step of subjecting the composite material to heat treatment to form a copper-molybdenum composite, wherein the heat treatment is a process in which the composite material is passed through a heat treatment furnace while being fed at a feed rate of 100 mm / min or less, the interior of the heat treatment furnace being a reducing atmosphere, and the composite material is held inside the furnace at a temperature of 700°C or higher and 1000°C or lower for 0.3 hours or longer.

11. The method for producing a copper-molybdenum composite according to claim 10, wherein the copper-molybdenum composite is a plate-like body having a first surface and a second surface, the copper content is 20 mass % or more and 80 mass % or less, and the standard deviation of the in-plane distribution of the nickel content measured at multiple locations on the first surface and the second surface is 5 ppm or less.

12. The method for producing a copper-molybdenum composite according to claim 11, wherein the maximum nickel content is 20 ppm or less.

13. A method for producing a copper-molybdenum composite according to any one of claims 10 to 12, wherein the copper-molybdenum composite is a plate-shaped body with a long side of 100 mm or more, a short side of 100 mm or more, and a thickness of 2 mm or more.

14. A method for manufacturing a heat sink, comprising the steps of preparing a copper-molybdenum composite by the method for manufacturing a copper-molybdenum composite according to any one of claims 10 to 13, and rolling the copper-molybdenum composite at a processing rate of 20% or more to form a heat sink, wherein the heat sink has a thermal conductivity in the thickness direction at room temperature of 150 W / m·K or more and 320 W / m·K or less, and a linear expansion coefficient from room temperature to 800°C in an in-plane direction perpendicular to the thickness direction of the heat sink of 6.5 ppm / K or more and 14.0 ppm / K or less.

15. The method for manufacturing a heat sink according to claim 14, further comprising the step of forming a metal coating on at least a portion of the first surface or the second surface, or on at least a portion of both the first surface and the second surface, of the heat sink, wherein the metal coating is made of at least one metal selected from the group consisting of nickel, nickel-phosphorus, copper, silver, gold, and platinum, or an alloy thereof.

16. A method for manufacturing a heat sink, comprising a step of alternately laminating, in the thickness direction, at least one first material layer made of the copper-molybdenum composite according to any one of claims 10 to 13 and a second material layer containing copper as the main component.

17. A method for manufacturing a heat sink according to any one of claims 14 to 16, further comprising a step of cutting into small pieces each having a long side of less than 100 mm and a short side of less than 100 mm.

Images