HEAT DISSIPATION SUBSTRATE AND METHOD FOR PRODUCING HEAT DISSIPATION SUBSTRATE

Abstract

A heat dissipation substrate having the maximum value of the coefficient of linear expansion of 10 ppm/K or less in any direction in a plane parallel to the surface within a temperature range from room temperature to 800° C. as well as a thermal conductivity of 250 W/m.Math.K or higher at 200° C. is produced by densifying an alloy composite of CuMo or CuW composed of Cu and coarse powder of Mo or W and subsequently cross-rolling the same alloy composite.

Claims

1. A production method for a heat dissipation substrate, comprising steps of: creating an alloy composite using Cu and particles of Mo or W as main components; densifying the alloy composite; and cross-rolling the alloy composite after the densifying step.

2. The production method for a heat dissipation substrate according to claim 1, wherein: the densified alloy composite are solid-phase sintered before the cross-rolling step.

3. A production method for a heat dissipation substrate, characterized in that: a heat dissipation substrate having a maximum value of a coefficient of linear expansion of 10 ppm/K or less in any direction in a plane parallel to a surface within a temperature range from RT to 800° C. as well as a thermal conductivity of 250 W/m.Math.K or higher at 200° C. is created by steps of: creating an alloy composite of CuMo or CuW using Cu and particles of Mo or W as main components, with at least 90% of the particles having a size within a range from 15 μm to 200 μm; densifying the alloy composite to increase a relative density of the alloy composite; solid-phase sintering the alloy composite after the densifying step; and cross-rolling the alloy composite after the solid-phase sintering step.

4. The production method for a heat dissipation substrate according to claim 1, wherein: the densifying step is performed by rolling the alloy composite to increase the relative density of the alloy composite to 99% or a higher level.

5. The production method for a heat dissipation substrate according to claim 4, wherein: the rolling is performed on the alloy composite in a canned and deaerated state.

6. The production method for a heat dissipation substrate according to claim 1, wherein: a metallic plating process is performed on the densified alloy composite before the cross-rolling step.

7. The production method for a heat dissipation substrate according to claim 1, wherein: the cross-rolling is a warm, hot or cold cross-rolling process or a combination of these kinds of cross-rolling.

8. A heat dissipation substrate including, as a main body, an alloy composite using Cu and particles of Mo or W as main components, wherein: a maximum coefficient of linear expansion in any direction in a plane parallel to the surface within a temperature range from 25° C. to 800° C. is equal to or less than 10 ppm/K, and a thermal conductivity at 200° C. is equal to or higher than 250 W/m.Math.K.

9. The heat dissipation substrate according to claim 8, wherein: the particles of Mo or W distributed inside the heat dissipation substrate have a flat shape spread in a plane parallel to a surface of the heat dissipation substrate, with at least 90% of the particles of Mo or W having a surface area within a range from 4.9×10.sup.−9m.sup.2 to 1.8×10.sup.−6m.sup.2.

10. The heat dissipation substrate according to claim 8, wherein: a metallic layer having a thickness of 1 μm or greater is formed on a surface of the alloy composite.

11. The heat dissipation substrate according to claim 8, wherein: one or a plurality of metallic layers are formed on each of obverse and reverse surfaces of the alloy composite.

12. A semiconductor package, comprising the heat dissipation substrate according to claim 8.

13. A semiconductor module, comprising the heat dissipation substrate according to claim 8.

14. A semiconductor module, comprising the heat dissipation substrate according to claim 8 having a surface on which a solder joint with a void percentage of equal to or lower than 5% is formed via a Ni-based plating.

15. The production method for a heat dissipation substrate according to claim 3, wherein: the densifying step is performed by rolling the alloy composite to increase the relative density of the alloy composite to 99% or a higher level.

16. The production method for a heat dissipation substrate according to claim 15, wherein: the rolling is performed on the alloy composite in a canned and deaerated state.

17. The production method for a heat dissipation substrate according to claim 3, wherein: a metallic plating process is performed on the densified alloy composite before the cross-rolling step.

18. The production method for a heat dissipation substrate according to claim 3, wherein: the cross-rolling is a warm, hot or cold cross-rolling process or a combination of these kinds of cross-rolling.

Description

BRIEF DESCRIPTION OF DRAWINGS

[0076] FIG. 1 is a graph showing the relationship between the temperature and the coefficient of linear expansion of representative heat dissipation substrates of CuW and CuMo.

[0077] FIG. 2 is a cross sectional view of a structure canned in a SUS case.

DESCRIPTION OF EMBODIMENTS

(Raw Material)

[0078] By using CuMo or CuW prepared using coarse powder of Mo or W, a heat dissipation substrate having a high thermal conductivity can be created. In the present embodiment, at least 90% of the particles of Mo or W need to have a particle size within a range from 15 um to 200 um. In other words, up to 10% of the entire amount of powder may have a particle size outside this range. If more than 10% of the particles have a size of 15 um or less, it is impossible to achieve the state where the coefficient of linear expansion has an appropriate value of 10 ppm/K or less and the thermal conductivity is 250 W/m.Math.K or higher at 200° C. If more than 10% of the particles have a size of 200 um or larger, only a minor improvement in the thermal conductivity can be achieved, while the powder price becomes considerably high. As for the Cu powder, there is no specific requirement, although electrolytic copper powder having a particle size within a range from 5μm to 10 um is preferable.

(Composition)

[0079] Both CuMo and CuW may have any composition as long as the obtained material has (1) a suitable coefficient of linear expansion for semiconductor modules and (2) a high thermal conductivity. Mixture of W and Mo is also permissible as long as the required characteristics concerning the coefficient of linear expansion and thermal conductivity are satisfied.

[0080] As for the additive metal, it has already been reported that adding an appropriate kind of metal improves the performance of the liquid metal infiltration or sintering process. There is no specific requirements concerning the kind and quantity of metallic element to be added as long as the obtained material has (1) a suitable coefficient of linear expansion for semiconductor modules and (2) a high thermal conductivity. However, the addition of metal is actually unrecommendable, since some kind of metal decreases the thermal conductivity. Accordingly, in the present embodiment, a high level of thermal conductivity is achieved without using additive metal, although this increases the difficulty of the creation of the alloy composite.

(Alloy Composite)

[0081] In the case of using Cu and coarse powder of Mo or W, either the liquid metal infiltration or sintering process may be used to create CuMo or CuW. Whichever method is used, there will be no significant difference in the characteristics or the like as long as an alloy composite having a relative density of 99% or higher after the rolling process is obtained using powder of Mo or W having almost the same particle size. The more economical method can be chosen.

(Densification)

[0082] A dense alloy composite having a high relative density is needed to obtain a heat dissipation substrate by cross-rolling. The densification may be achieved by any method. Densification of CuMo or CuW to a relative density of 99% or higher normally requires high temperature and high pressure. For example, such methods as hot pressing or forging can be used, although these methods uneconomically require a large system. Hot forging is also unfavorable since it causes oxidization of Cu, Mo or W on the surface layer as well as in the inner region of the alloy composite.

[0083] By comparison, densifying the alloy composite by a hot rolling process followed by solid-phase sintering is an effective method, since the subsequent manufacturing process is also a rolling process (i.e. a cross-rolling process, which will be described later). However, if the alloy composite has a low relative density, a measure for preventing its oxidization is required, otherwise its surface layer or inner region will undergo oxidization during the rolling process. A solution to this problem is to contain the alloy composite in a SUS case for the prevention of oxidization and circumferential cracking, and deaerate and roll this case, whereby an alloy composite which is densified to a relative density of 99% or higher and suitable for the subsequent cross-rolling process can be obtained. The process of densifying the alloy composite to a relative density of 99% or higher can be controlled by optimizing the processing conditions by a preliminary experiment. The canning of the alloy composite minimizes circumferential breakage or cracking, whereby the yield of the cross-rolling process is improved. The solid-phase sintering of this alloy composite at a temperature equal to or lower than the melting point of Cu in hydrogen atmosphere repairs the separation of the particle surfaces of Mo or W and Cu as well as reduces oxides which occurs due to the residual oxygen. Thus, a suitable alloy composite for rolling is obtained. A preferable condition of the solid-phase sintering is to hold the alloy composite in hydrogen atmosphere at a temperature equal to or higher than 800° C. and lower than the melting point of Cu (or lower than the melting points of all kinds of metal used as the main components of the alloy composite) for 60 minutes. Such a solid-phase sintering allows for a satisfactory rolling process, with the result that a dense heat dissipation substrate which does not allow blisters or similar defects of the alloy composite to occur even under a high temperature of 800° C. for silver soldering can be obtained.

[0084] Another possible method for obtaining a suitable alloy composite for cross-rolling is to use an alloy composite with a low relative density and repeatedly perform a cross-rolling process with a low percentage of rolling reduction and a solid-phase sintering process until the relative density becomes 99% or higher. However, this method is time-consuming as well as uneconomical.

(Cu-Plating on Surface Layer)

[0085] As in the case of CuMo or CuW having a Mo content of 50% or lower or W content of 60% or lower with the balance being Cu, if the Cu content is considerably high, it is not always necessary to plate the surface layer with Cu in the rolling process. However, with a decrease in the Cu content, the area where the particles of Mo or W are in contact with or overlap each other increases, causing such phenomena as the detachment of the particles of Mo or W, or formation of burrs, during the rolling process. Such a problem can be addressed by plating the surface layer with Cu before the rolling process. From an economical point of view, the plating thickness should preferably be equal to or less than 10 μm. However, the plating may become ineffective if its thickness is as small as 3 μm or even thinner. Although the plating layer becomes thinner through the rolling process, the final Ni-plating can be satisfactorily performed if the Cu layer which eventually remains has a thickness of approximately 1 μm over the entire surface.

[0086] It is also possible to increase the thickness of the Cu plating so as to create a clad structure, similar to Cu/CuMo/Cu or Cu/CuW/Cu. A clad structure is a structure in which one or a plurality of metallic layers are formed on each of the obverse and reverse surfaces of an alloy composite. The use of the heat dissipation substrate having such a clad structure improves the compatibility with the Ni-based plating process which is performed in the final processing of the heat dissipation substrate (i.e. the degree of adhesion of the Ni-based plating), thus enabling the production of a heat dissipation substrate having a high-quality Ni-based plating.

(Cross-Rolling)

[0087] In the cross-rolling process, the alloy composite placed in a non-oxidizing or reducing atmosphere and heated to 300° C. or a higher temperature is alternately rolled in the

[0088] X-axis and Y-axis directions (where both X and Y axes are defined in a plane parallel to the surface, with the thickness direction defined as the Z axis). The cross-rolling decreases and stabilizes the maximum coefficient of linear expansion within a temperature range from RT to 800° C. in any direction in the plane parallel to the surface (i.e. not only the X-axis and Y-axis directions in which the cross-rolling is performed, but also in other directions in the plane) as well as improves and stabilizes the thermal conductivity. A simple mono-axial rolling is not suitable for a heat dissipation substrate, since a considerable difference in the coefficient of linear expansion occurs between the direction in which the cross-rolling is performed (e.g. X-axis direction) and the perpendicular direction (Y-axis direction). It is preferable to alternately perform the cross-rolling in the X-axis and Y-axis directions. By such a cross-rolling process, the particles of Mo or W distributed inside the alloy composite are shaped into a disk-like flat form in a plane parallel to the surface of the heat dissipation substrate. The percentage of rolling reduction of the alloy composite (i.e. the percentage of rolling reduction by the two processes of densifying and cross-rolling) at this stage is within a range from 50% to 80%. As already noted, at least 90% of the particles of Mo or W have a particle size within a range from 15 μm to 200 μm. Accordingly, with the shape of each particle of Mo or W approximated by a sphere (with a volume of 4/3π.sup.3, where r is the radius of the sphere) and that of each particle after the cross-rolling (percentage of rolling reduction, P) approximated by a disk-like flat body (with a volume of r×(1−P)×πr′.sup.3, where r′ is the radius of the circular bottom face of the disk-like flat body after the cross-rolling), the size of the particles in a plane parallel to the surface of the heat dissipation substrate after the cross-rolling will be within a range from approximately 17 μm (which is the size in the case of using spherical particles with a radius of 15 μm as the raw material and performing the cross-rolling process with a percentage of rolling reduction of 50%) to approximately 366 μm (which is the size in the case of using spherical particles with a radius of 200 μm as the raw material and performing the cross-rolling process with a percentage of rolling reduction of 80%).

[0089] According to past performance records, there will be no problem in practical use if the difference in the coefficient of linear expansion between the X-axis and Y-axis directions is equal to or less than 20%, whereas a difference greater than 20% puts some restrictions on the use. By an appropriate selection of the raw materials and their composition as well as the shape of the used powder of Mo or W, along with an optimization of the cross-rolling conditions, a heat dissipation substrate having the required characteristics can be obtained.

[0090] The cross-rolling may be performed in any order in the X-axis and Y-axis directions, and any number of times, as long as the difference in the coefficient of linear expansion between the X-axis and Y-axis directions of the obtained heat dissipation substrate becomes equal to or less than 20%. In the present embodiment, the rolling is performed in two orthogonal directions (X-axis and Y-axis directions). However, the objective of the cross-rolling is to achieve a coefficient of linear expansion of 10 ppm/K or less in any direction in a plane parallel to the surface as well as decrease the anisotropy in the coefficient of linear expansion. As long as this objective is achieved, the cross-rolling does not always need to be performed in two orthogonal directions but may be performed in any two or more non-parallel directions (i.e. two or more mutually intersecting directions).

[0091] In an alloy composite with a relative density of 99% or higher, if its thickness is decreased to one fifth or less of the thickness before the rolling process, the flattened Mo or W may possibly be split into pieces, causing a variation in the coefficient of linear expansion and thermal conductivity. Accordingly, the rolling process should preferably be discontinued at a thickness greater than one fifth of the thickness of the alloy composite before the rolling process.

[0092] The rolling process may be any of the cold, warm and hot rolling processes. Cold rolling is not productive since a high percentage of rolling reduction cannot be achieved. For

[0093] CuMo, a warm rolling process performed around 400° C. is preferable. For CuW, a hot rolling process performed around 600° C. is preferable. The quality of the rolled product can be improved by performing an acid cleaning, reduction treatment, buffing or similar process for each rolling operation in order to remove oxides from the surface layer. By performing a cold rolling process after a thermal treatment in hydrogen atmosphere, a finished product which has a smooth surface and is suitable for a heat dissipation substrate can be obtained.

(Final Plating)

[0094] Although Mo and W are not always easy to be plated with metal, the Ni-based final plating is performed in order to prevent Cu in CuMo or CuW from being eroded in the silver-soldering or soft-soldering process. For high-grade products, Au-plating may additionally be performed on the Ni-based final plating in order to improve the quality of the soldering of the semiconductor device as well as enhance the commercial value. It should be noted that the term “Ni-based plating” means plating with Ni or Ni alloy.

[0095] For a heat dissipation substrate made of Cu, a Ni-based one-time direct plating process with no thermal treatment is sufficient. In the case of CuMo or CuW, since the area where Mo or W is exposed cannot be easily plated with metal, a multilayer plating process needs to be performed, such as thermal treatment+Ni-plating+thermal treatment+Ni-plating. Such a process requires a long period of time, causing a long delivery time and an increased production cost. In the case of the heat dissipation substrate of the present embodiment, although the multilayer plating process can be similarly performed, a Ni-based one-time direct plating process may be sufficient in the case where the Cu-plating layer formed before the rolling process still remains.

(Other Remarks)

[0096] In a semiconductor module, the quality of the solder joint portion between the heat dissipation substrate and the semiconductor device is essential. The void percentage in this portion must meet a strict condition. Most commonly used solder materials for semiconductor devices are AuSn (melting point, 280° C.) and AuSi (melting point, 363° C.), both of which are in conformity with the requirements of the Pb-free production and high-temperature operation. For a semiconductor device which needs to withstand 200° C. or higher temperatures, an even higher quality is desired. In such a case, the device may be soldered onto an Au-plated heat dissipation substrate. Ni-based final plating methods which are suited for Cu, CuMo and CuW have already been developed. In the present invention, if a Cu plating layer is present, a 3-μm Ni-B final plating process can be directly performed, and its quality can be controlled by a blister test. Meanwhile, a multilayer Ni-based final plating as in the conventional CuMo or CuW is also frequently desired. In such a case, the blister test can be similarly used for the check and control of the quality. It has been commonly known that a product which has passed a blister test will not cause any problem concerning the silver soldering, solder joint or practical use.

<Evaluation of Heat Dissipation Substrate>

(Measurement of Coefficient of Linear Expansion)

[0097] From each alloy composite obtained by the previously described cross-rolling process, a sample measuring 10 mm in the X-axis direction, 4 mm in the Y-axis direction and 2-2.5 mm in thickness (Z-axis direction) was cut out by wire electrical discharge machining (hereinafter abbreviated as the “WEDM”). Using a device for measuring coefficient of linear expansion (manufactured by Seiko Instruments Inc.), the coefficient of linear expansion within a temperature range from RT to 800° C. was measured in both X-axis and Y-axis directions, and the larger value was adopted.

(Measurement of Thermal Conductivity)

[0098] From each alloy composite obtained by the previously described cross-rolling process, a sample measuring 10 mm in diameter and 2-2.5 mm in thickness was cut out by WEDM. Using a laser-flash thermal conductivity meter (TC-7000, manufactured by Advance Riko, Inc.), the thermal conductivity was measured at 200° C. in hydrogen atmosphere.

(Blister Test on Plating)

[0099] A multilayer Ni-plating process and single-layer direct plating process were performed on samples measuring 5 mm×25 mm. After those samples were held in the air at 400° C. for 30 minutes, the appearance of each sample was observed with a stereoscopic microscope at 10-fold magnification. Samples which had no blister on the metallic plating layer were judged to be “OK”, while samples on which a blister was recognized were judged to be “NG”, regardless of the size of the blister.

EXAMPLE

Example 1: CuMo with 40 wt % Cu; Liquid Metal Infiltration, Densification and Rolling; Sample No. 6

[0100] Mo powder with an average particle size of 60 μm was mixed with 3 wt % of electrolytic copper powder with an average particle size of 10 μm and 1 wt % of paraffin wax. The obtained mixed powder was press-molded in a 50-mm×50-mm mold. The molded object was heated at 600° C. for 60 minutes in hydrogen atmosphere to remove wax. The same object was further heated to 1000° C. in hydrogen atmosphere to obtain a skeleton. A Cu plate was placed on this skeleton and heated at 1250° C. for 60 minutes in hydrogen atmosphere to infiltrate molten Cu into the skeleton. In this manner, a CuMo alloy composite measuring 50 mm×50 mm×6 mm with 40 wt % Cu was created. Residues of the infiltration Cu on surface layer of the alloy composite as well as defects on the surface layer were removed by cutting work. After the alloy composite was contained in a SUS case and deaerated, the ends of the case were welded to complete the canning. The canned alloy composite was cross-rolled at 800° C. After the relative density of the alloy composite reached a level of 99% or higher, the alloy composite was removed from the case and solid-phase sintered at 950° C. for 60 minutes in hydrogen atmosphere. The solid-phase-sintered (or densified) alloy composite was subjected to a 10-μm Cu-plating process, and subsequently warm-cross-rolled at 400° C. until the thickness was reduced to 2 mm. That is to say, the percentage of rolling reduction of the two-stage cross-rolling process was 66.6% (=(6 mm-2 mm)/6 mm).

[0101] After an additional thermal treatment was performed at 450° C. for 15 minutes in hydrogen atmosphere, the alloy composite was cold-rolled to smooth its surface.

[0102] Some of the heat dissipation substrates obtained in this manner were subjected to the multilayer Ni-based plating process, while others were subjected to the single-layer direct Ni-plating process. The blister test was performed on both groups.

[0103] The coefficient of linear expansion and thermal conductivity of each sample were also measured.

[0104] The result is shown in Table 2.

[0105] [0068]

Example 2: CuMo with 40 wt % Cu; Sintering, Densification and Rolling; Sample No. 7

[0106] Mo powder with an average particle size of 60 μm was mixed with electrolytic copper powder with an average particle size of 10 μm to prepare a mixed powder having a Cu content of 40 wt %, with the balance being Mo. The obtained mixed powder was press-molded in a 50-mm×50-mm mold. The molded object was liquid-phase sintered at 1250° C. for 60 minutes in hydrogen atmosphere to obtain a CuMo alloy composite measuring 50 mm x 50 mm×6 mm. Defects on the surface layer of the alloy composite were removed by cutting work. After the alloy composite was contained in a SUS case and deaerated, the ends of the case were welded to complete the canning. The canned alloy composite was cross-rolled at 800° C. After the relative density of the alloy composite reached a level of 99% or higher, the alloy composite was removed from the case and solid-phase sintered at 950° C. for 60 minutes in hydrogen atmosphere. The sintered alloy composite was subjected to a 10-μm Cu-plating process and subsequently cross-rolled at 400° C. to obtain a plate material with a thickness of 2 mm. That is to say, the percentage of rolling reduction of the two-stage cross-rolling process was 66.6% (=(6 mm-2 mm)/6 mm).

[0107] The plate material was thermally treated at 450° C. for 15 minutes in hydrogen atmosphere and subsequently cold-rolled to smooth its surface.

[0108] Some of the heat dissipation substrates obtained in this manner were subjected to the multilayer Ni-based plating process, while others were subjected to the single-layer direct plating process. The blister test was performed on both groups.

[0109] The coefficient of linear expansion and thermal conductivity of each sample were also measured.

[0110] The result is shown in Table 2.

Example 3: CuW with 45 wt % Cu; Sintering and Rolling; Sample No. 20

[0111] Mo powder with an average particle size of 60 μm was mixed with electrolytic copper powder with an average particle size of 10 μm to prepare a mixed powder having a Cu content of 45 wt %, with the balance being Mo. The obtained mixed powder was press-molded in a 50-mm×50-mm mold. The molded object was liquid-phase sintered at 1250° C. for 60 minutes in hydrogen atmosphere to obtain a CuMo alloy composite measuring 50 mm×50 mm×6 mm.

[0112] Defects on the surface layer of the alloy composite were removed by cutting work. After the alloy composite was contained in a SUS case and deaerated, the ends of the case were welded to complete the canning. The canned alloy composite was cross-rolled at 800° C. After the relative density of the alloy composite reached a level of 99% or higher, the alloy composite was removed from the case and solid-phase sintered at 1000° C. for 60 minutes in hydrogen atmosphere. The sintered alloy composite was subjected to a 10-μm Cu-plating process and subsequently cross-rolled at 600° C. until the thickness was reduced to 2 mm. That is to say, the percentage of rolling reduction of the two-stage cross-rolling process was 66.6% (=(6 mm-2 mm)/6 mm).

[0113] Some of the heat dissipation substrates obtained in this manner were subjected to the multilayer Ni-based plating process, while others were subjected to the single-layer direct plating process. The blister test was performed on both groups.

[0114] The coefficient of linear expansion and thermal conductivity of each sample were also measured.

[0115] The result is shown in Table 2.

Example 4: Evaluation of a Semiconductor Module having a Semiconductor Device Mounted on a Heat Dissipation Substrate in a Package

[0116] Using the heat dissipation substrate of Example 2 having a coefficient of linear expansion of 9.1 ppm/K and thermal conductivity of 293 W/m.Math.K, a package was created by silver-soldering members made of ceramic, Kovar and other materials onto the heat dissipation substrate in hydrogen atmosphere at 800° C. It was checked that neither separation nor cracking was present on the package. The metallic electrode layer of a Si device measuring 10 mm×10 mm×0.7 mm was joined onto this package, using a high-temperature AuSi solder (melting point, 363° C.) at 400° C., to obtain a semiconductor module. By an ultrasonic measurement, it was confirmed that the percentage of the void area in the solder joint portion was not higher than 3%. In general, when the final plating is a 3-μm Ni-B plating, the SnAgCu solder (melting point, 218° C.) needs to pass an extremely stringent evaluation test which requires the void percentage measured by ultrasonic measurement to be 5% or lower. It is commonly known that any material which satisfies this requirement causes no problem in silver soldering, other kinds of soldering, resin adhesion or the like. The voids formed in the soldering process reflect pinholes which are present on the surface of the heat dissipation layer before the Ni-based final plating process is performed. In other words, the evaluation condition for SnAgCu (melting point, 218° C.) can be satisfied by using a heat dissipation substrate on which the percentage of the pinholes (defects) is equal to or lower than 5%. In Example 4, the void percentage is not higher than 3%. Thus, all of the aforementioned requirements are satisfied.

[0117] A heat cycle test module (−40° C. to 225° C., 3000 times) was performed on the same semiconductor. Meanwhile, for comparison, another package was created using a conventional heat dissipation substrate of CuMo with 40 wt % Cu in the same size, with a coefficient of thermal expansion of 9.1 ppm/K (the same as in Example 2) and thermal conductivity of 213 W/m.Math.K. After the devices were mounted on this package, the heat cycle test (−40° C. to 225° C., 3000 times) was similarly performed.

[0118] The result confirmed that separation, cracking or other problems did not occur on any of the two samples.

(Interpretation of Present Disclosure-1)

[0119] By the present invention, a high-performance heat dissipation substrate which satisfies the requirements for use with future high-performance semiconductor modules can be obtained.

(Interpretation of Present Disclosure-2)

[0120] The present invention is not limited to the previous embodiment. Any mode of modification will also be included in the present invention as long as the objective of the present invention is thereby achieved. Specific structures, modes or the like for carrying out the present invention may also be altered as long as the objective of the present invention is thereby achieved.

[0121] The presently disclosed embodiment and examples should be considered, in all aspects, as mere examples of non-restrictive nature. The subject matter is as set forth in claims and should not be limited to the previous descriptions.

REFERENCE SIGNS LIST

[0122] 1 . . . Alloy Composite Created by Liquid Metal Infiltration or Sintering Process [0123] 2 . . . SUS Canning Case [0124] 3 . . . Circumferentially Welded Joint Portion