JP5203906B2 - Bi-containing solder foil manufacturing method, Bi-containing solder foil, joined body, and power semiconductor module - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of manufacturing Bi-containing solder foil and Bi-containing solder foil, and to provide a joined body that is highly heat resistant and accurately welded and a power semiconductor module. <P>SOLUTION: In the method of manufacturing Bi-containing solder foil, powder of a solder material containing bismuth (Bi) is formed into a sheet by a powder rolling method. The ratio of a long side and a short side in the powder of the solder material is suitably &ge;1.2 and &le;3.0 while the mean particle diameter of the powder is suitably &ge;5 &mu;m and &le;200 &mu;m. The rolling by the powder rolling method is desirably performed at &ge;-20&deg;C and &le;269&deg;C. Also, in the joined body and the power semiconductor module, the Bi-containing solder foil is used in the weld zone. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to a method for producing a Bi-containing solder foil, a Bi-containing solder foil, a joined body, and a power semiconductor module.

  In a large area and laminated mounting structure, such as a power module structure for an inverter of a hybrid vehicle, a sheet-like solder foil is preferably used so that an object to be joined such as an insulating substrate or a heat radiating substrate is uniformly bonded. It has been. Currently, in inverter bonding, for example, a solder sheet having a thickness of 100 μm and a width of 2 cm is wound in a roll shape, mounted on a bonding apparatus, cut out at a predetermined length, and placed at a predetermined position on a workpiece by a robot arm. Precisely programmed to place.

  Conventionally, lead (Pb) solder materials having excellent ductility have been widely used for sheet-like solder foils. In particular, a Pb—Sn solder material is widely used, and the melting point is changed in the range of about 183 to 300 ° C. by changing the ratio of lead (Pb) and tin (Sn) to obtain solder materials having various melting points ( For example, refer nonpatent literature 1.). However, since lead is toxic, it is being abandoned and development of a lead-free solder material is desired.

  Among the requirements for such solder materials, Sn-based solder materials such as Sn—Ag alloys and Sn—Cu alloys have been proposed as lead-free solder materials. However, since the melting point of the Sn-based solder material is about 220 ° C., the tensile strength is remarkably reduced around 200 ° C., and it is difficult to put it into practical use for a next-generation power semiconductor element that generates heat exceeding 200 ° C.

Therefore, bismuth (Bi) solder materials have been studied from the viewpoint that a melting point is about 270 ° C. and a bonded body having excellent heat resistance can be obtained.
On the other hand, it is known that Bi-based solder materials have low ductility and are brittle (see, for example, Non-Patent Document 2). In particular, it is vulnerable to shearing, and when strain due to impact or thermal stress accumulates, it tends to cause breakage starting from the strain.

As a measure for improving ductility, the addition of a third component has been studied for Sn-Bi solder materials. For example, a method of making the structure finer by adding Ag (see, for example, Patent Document 1), and a method of suppressing the transformation of Sn from a β phase to an α phase by adding Sb (for example, see Patent Document 2). , A method of adding In (see, for example, Patent Document 3), and the like.
Furthermore, it has been reported that Sn-Bi solder material containing 25 to 55% by mass of Bi and the balance being Sn and inevitable impurities is improved in ductility (see, for example, Patent Document 4).

  Further, in order to eliminate the brittleness of the Bi-based solder material, a solder material in which an alloy that causes thermoelastic martensitic transformation in Bi, specifically, a Cu-based alloy containing Mn and Al has been proposed (for example, , See Patent Document 5).

  However, the above technique is a method of changing the solder composition. Further, even with a Bi-based solder material whose ductility is improved by the above technique, the ductility is not improved until the ingot is rolled into a sheet and cannot be processed into a solder foil.

  Until now, solder materials with low workability have been soldered by screen printing a solder paste kneaded with flux. However, it is desirable not to leave a flux residue at the inverter junction, and it is desired not to use flux.

Solder materials that are difficult to process into a sheet can be soldered in the form of pellets or lumps, but it is extremely difficult to uniformly bond when the joint surface is large.
Thus, there has been a demand for a method of easily processing a solder material that is difficult to be processed into a sheet shape, in particular, a Bi-based solder material having a high melting point into a sheet shape.

  By the way, as a general sheeting method, there are a method of rolling an ingot and a method of forming a sheet by applying pressure after pulverizing the material into a powder. An example of this method is a powder rolling method.

  Specifically, for example, a method for producing a nickel brazing material sheet using a powder rolling method is disclosed (for example, see Patent Document 6). In addition, the powder rolling method has been proposed as a technique for producing a steel sheet (see, for example, Patent Documents 7 and 8).

Also, a metal ball such as copper (Cu) and a solder ball such as tin (Sn) are blended, the whole is uniformly compressed in advance, the solder ball such as Sn is plastically flowed between the metal balls, and the gap is soldered ( There has been proposed a method of rolling a composite molded body filled with (Sn-based solder after plastic deformation) (see, for example, Patent Documents 9 and 10).
In this method, aluminum (Al), gold (Au), silver (Ag), zinc (Zn) -aluminum (Al) alloy, gold (Au) -tin (Sn) alloy, etc. are used as metal balls in addition to Cu. Applicable.
JP-A-8-252688 Japanese Unexamined Patent Publication No. 7-40079 JP-A-8-150493 Japanese Patent Laid-Open No. 2000-141079 JP-A-2005-313230 JP 2004-82218 A Japanese Patent Laid-Open No. 6-025870 JP 2002-275601 A JP 2002-301588 A JP 2002-305213 A Yoichiro Baba “Efforts to Ensure HV Inverter Quality” Outline of the National Conference of the Japan Welding Society, Chapter 77 (2005-9) Journal of the Japan Institute of Metals, vol. 57 (1993), 455-462)

The subject of this invention is providing the manufacturing method of Bi containing solder foil, and Bi containing solder foil.
Another object of the present invention is to provide a bonded body and a power semiconductor module that have high heat resistance and are densely bonded.

  Bi-based solder materials are highly brittle, and even if the ingot is rolled as it is, it is not pulverized but integrated, and a sheet cannot be obtained. Therefore, a Bi-based solder material having high brittleness cannot be processed into a sheet shape simply by rolling.

Here, there is a powder rolling method as one of methods for processing into a sheet shape. However, the conventional powder rolling method is used for forming steel plates, nickel, and Cu / Sn, and is a method using a material having ductility. In the above-mentioned Patent Documents 9 and 10 for explaining Cu / Sn sheet formation, it is explained that ductile Sn is plastically flowed between Cu to form a sheet, and as is clear from this description, the powder rolling method is used. Then, it was common sense to use one or more ductile materials.
In addition, in the research process of the inventors leading to the present invention, it was found that silicon (Si) materials such as glass and ceramics, which are brittle metal materials, are difficult to form into a sheet even when the powder rolling method is applied. did.

From the above fact, it was thought that the brittle material could not be made into a sheet even by adopting the powder rolling method, but by further diligent research by the present inventors, by devising the shape and the like of the Bi-based material, It has been found that Bi-based solder sheets can be formed by powder rolling.
As described above, the inventors' diligent research has revealed that the problems of the present invention are solved by the following invention.

That is, the invention described in claim 1
A method for producing a Bi-containing solder foil , comprising mixing Bi powder and CuAlMn alloy powder, and forming a powder of a solder material containing bismuth (Bi) containing 30% by mass or more of the Bi powder by a powder rolling method. It is.

The invention described in claim 2
A method for producing a Bi-containing solder foil comprising mixing a Bi powder and a Cu powder, and forming a powder of a solder material containing bismuth (Bi) containing 30% by mass or more of the Bi powder by a powder rolling method It is.

The invention according to claim 3
3. The method for producing a Bi-containing solder foil according to claim 1, wherein an average particle diameter of the Bi powder is 5 μm or more and 200 μm or less.

The invention according to claim 4
The ratio of the long side to the short side of each powder particle in the Bi powder is in the range of 1.2 or more and 3.0 or less, according to any one of claims 1 to 3. It is a manufacturing method of Bi containing solder foil.

The invention described in claim 5
The rolling by powder rolling method is a method for producing a Bi-containing solder foil according to any one of claims 1 to 4, characterized in that it is carried out at -20 ° C. or higher 269 ° C. or less.

The invention described in claim 6
A Bi-containing solder foil comprising a powder of a solder material containing 30% by mass or more of Bi powder and further containing CuAlMn alloy powder or Cu powder containing bismuth (Bi) by a powder rolling method.

The invention described in claim 7
The Bi-containing solder foil according to claim 6 , wherein an average thickness is 30 μm or more and 250 μm or less.

The invention according to claim 8 provides:
A first member and a second member;
Between the said 1st member and the 2nd member, it is the joined body formed by joining with the Bi containing solder foil of Claim 6 or Claim 7 .

The invention according to claim 9 is:
A power semiconductor element, an insulating substrate, and a heat sink;
Bi content of Claim 6 or Claim 7 in which at least one of the 1st junction part between the said power semiconductor element and the said insulation board | substrate and the 2nd junction part between the said insulation board | substrate and the said heat sink is contained. It is a power semiconductor module characterized by being joined by solder foil.

The invention according to claim 10 is:
The power semiconductor module according to claim 9 , wherein the power semiconductor element is made of GaN or SiC.

The invention according to claim 11
Wherein the insulating substrate is a Si ₃ N ₄ layers, power of claim 9 or claim 10, characterized in that it comprises a conductive layer formed of copper (Cu) on both the surface of the Si ₃ N ₄ layer semiconductor It is a module.

The invention according to claim 12
The heat radiating plate, any one of claims 9 to 11 which is a laminate of Cu layer / Mo layer / Cu layer having a double-sided copper (Cu) layer of molybdenum (Mo) layer It is a power semiconductor module of description.

The invention according to claim 13
13. The power semiconductor module according to claim 12 , wherein a ratio of a thickness of Cu layer / Mo layer / Cu layer in the heat radiating plate is 1/5/1 to 1/12/1.

  ADVANTAGE OF THE INVENTION According to this invention, the manufacturing method of Bi containing solder foil and Bi containing solder foil can be provided. In addition, according to the present invention, it is possible to provide a bonded body and a power semiconductor module that have high heat resistance and are densely bonded.

  Hereinafter, with reference to the drawings, a Bi-containing solder foil manufacturing method of the present invention, a Bi-containing solder foil obtained by the method, a joined body joined with the Bi-containing solder foil, and a power semiconductor module as an example of the joined body explain.

[Production Method of First Embodiment]
In the first embodiment, a method for producing a pure Bi solder foil will be described. In addition, the pure Bi solder foil here does not mean only 100% pure Bi solder foil, but may contain impurities contained in the raw material.

FIG. 1 is a schematic view of a solder foil manufacturing apparatus (powder rolling machine) used in the present embodiment. In FIG. 1, reference numeral 2 is a pure Bi solder foil, 4 is Bi powder, and 6 and 7 are rolling rolls.
The solder foil manufacturing apparatus (powder rolling mill) is composed of a pair of rolling rolls 6 and 7, as shown in FIG. Although not shown, a heating device may be provided. The pair of rolling rolls 6 and 7 are arranged so that their peripheral surfaces face each other at a predetermined interval.

  The Bi powder 4 is supplied between the pair of rolling rolls 6 and 7 from above, and is formed between the rolling rolls 6 and 7 by rotating the rolling rolls 6 and 7 in the arrow X direction. Sequentially fed into the gap. Then, the Bi powder 4 is sequentially fed in the lower Y direction while being pressed by the rolling rolls 6 and 7 and formed into a sheet shape to form the pure Bi solder foil 2.

  The Bi-containing solder material is highly brittle, and even if the ingot is rolled as it is, it is only pulverized and not integrated, but the reason why it is integrated when it is rolled from the powder shape is estimated as follows. However, the present invention is not limited by the estimation.

  When an ingot of a Bi-containing solder material is rolled, each crystal grain present in the ingot has low brittleness but tends to be deformed by rolling. However, before rolling and deforming to a uniform thickness because there are few slip surfaces, Many cracks are generated between crystal grains. Therefore, a relatively large Bi-containing solder ingot breaks at a rolling rate of several percent. However, when powder rolling is performed with a Bi-containing solder material in a powder state, deformation of each powder crystal due to rolling occurs within a brittle range, and the recrystallization temperature of Bi is around room temperature. It is presumed that they are easily adhered and deformed, and as a result, they are integrated into a sheet.

  Moreover, the Si material which is a brittle material cannot be formed into a sheet by the powder rolling method, whereas the reason why the Bi-containing solder material can be formed into a sheet by the powder rolling method is due to the difference in melting point of each material with the recrystallization temperature. I guess it. In the powder rolling method, it is presumed that the powder is integrated when rolled at a temperature higher than the recrystallization temperature. However, the recrystallization temperature of the Si material has a melting point of Si of 1414 ° C. (1687 K). The crystallization temperature is about 450 ° C. (840 K), and it is estimated that it is difficult to integrate by rolling the powder around room temperature. However, the present invention is not limited by the guess.

The average particle diameter of the Bi powder 4 is preferably smaller than the target sheet thickness, and more preferably 200 μm or less. The average particle size is preferably 5 μm or more. More preferably, the average particle diameter of the Bi powder 4 is not less than 10 μm and not more than 150 μm, and more preferably not less than 20 μm and not more than 100 μm.
When the average particle size is within the above range, an appropriate pressure is applied between the rolling rolls 6 and 7, and the control of brittle properties is eased.
In addition, the average particle diameter of Bi powder 4 means the arithmetic mean value when 100 powders are measured with an optical microscope.

The Bi powder 4 may have any shape such as a spherical shape, a rod shape, and an indefinite shape, but a rod-shaped powder is preferable because the pressure from the rolling rolls 6 and 7 is efficiently applied.
The ratio of the long side to the short side in the rod-shaped powder is preferably 1.2 or more and 3.0 or less, more preferably 1.3 or more and 2.8 or less, and 1.5 or more and 2.5 or less. More preferably. When the ratio is within the above range, the deformation and adhesion of each rod during rolling are promoted by rolling, and an increase in porosity can be suppressed.

  Here, the long side and the short side of the rod-shaped powder are values determined by measuring 100 powders with an optical microscope. When the shape of the rod-like powder is approximated to a rectangular parallelepiped and the side lengths are a, b, and c, the shortest of a, b, and c is the short side, and the longest is the long side. .

In addition, the average particle diameter in rod-shaped powder means the value calculated | required by calculating | requiring the average value of each of the said side length a, b, and c from 100 powders, and a following formula.
Average particle diameter of rod-shaped powder = (average value of side length a + average value of side length b + average value of side length c) / 3

  Bi powder 4 can be produced by a known method. For example, pulverization with a mortar or a roughing machine, and further a powder production technique such as atomization can be applied.

  The lot-like Bi powder 4 can be produced by finely crushing an ingot-like or needle-like Bi raw material with a hammer or the like, and then crushing for about 30 minutes with a mortar or a rough machine.

  The powder produced by pulverization or atomization is preferably classified so that pressure is uniformly applied during rolling. A known method can be applied to the classification, and either a dry method or a wet method may be used. There are methods such as air classification and electrostatic classification as the dry method, and there are methods such as sedimentation as the wet method.

The temperature during rolling is preferably −20 ° C. or higher and 269 ° C. or lower, more preferably −5 ° C. or higher and 100 ° C. or lower, and further preferably 0 ° C. or higher and 50 ° C. or lower. If it is less than −20 ° C., it is much lower than the recrystallization temperature of Bi (about 0 ° C.), and if it is 270 ° C. or more, it is more than the melting point of Bi, so it is difficult to form a sheet. The temperature at the time of rolling takes into consideration the heat generated by the powder being subjected to rolling deformation and the heat generated by the contact between the roll and the powder.
If it is the said temperature range, a heating apparatus (not shown) may be provided in the rolling rolls 6 and 7, and you may heat after passing the rolling rolls 6 and 7.

[Production Method of Second Embodiment]
In the manufacturing method of the second embodiment, a manufacturing method of a solder foil (hereinafter referred to as “Bi-based solder foil”) containing other metals or alloys in addition to Bi will be described.
FIG. 2 is a diagram for explaining the manufacturing method of the second embodiment. In FIG. 2, 4 is Bi powder, 5 is other metal powder or alloy powder, 6 and 7 are rolling rolls, and 8 is Bi-based solder foil.
The manufacturing method according to the second embodiment differs from the manufacturing method according to the first embodiment in that other metal powder or alloy powder 5 is used together with the Bi powder 4 described above, but the manufacturing method according to the first embodiment is otherwise. It is the same as the method.

The Bi powder 4 and the other metal powder or alloy powder 5 are sequentially fed in the lower Y direction while being pressed by the rolling rolls 6 and 7 and formed into a sheet shape, thereby forming a Bi-based solder foil 8.
In addition, in the manufacturing method of Bi type solder foil of this embodiment, in addition to the advantage that a brittle Bi containing material can be made into a sheet, it has the advantage that a manufacturing process can be simplified.

In the production of the conventional Bi-based solder material, other metal or alloy powder (CuAlMn alloy powder, Cu powder, etc.) to be added to Bi is prepared and dispersed in Bi to produce an ingot of a solder material containing Bi. It was. When making an ingot, complicated adjustments, such as improving the dispersibility of the other metal powder or alloy powder in Bi, were required. However, according to the present embodiment, a Bi-based solder foil can be obtained without producing an ingot in advance, so that the production process can be simplified.
Specifically, in the manufacturing method of the Bi-based solder foil of this embodiment, other metal powder or alloy powder 5 and Bi powder 4 may be prepared, and these powders may be supplied to the powder rolling apparatus at a desired ratio.

The composition of the Bi-based solder foil 8 is not particularly limited as long as it contains Bi. In order to make use of the characteristics of Bi (for example, high melting point), it is preferable to contain 30% by mass or more of Bi in the Bi-based solder foil. The composition of the Bi-based solder foil 8 can be appropriately changed according to the application of the solder.
Examples of the Bi-based solder foil 8 include those obtained by adding Cu, Ni, and Ag to Bi. From the viewpoint of eliminating the brittleness of Bi to increase the mechanical strength and not lowering the solidus temperature. The Bi-based solder foils described in (1) to (3) below are preferable. Furthermore, from the viewpoint of increasing the liquidus temperature, the following (1) or (2) Bi solder foil is preferable.

(1) Bi-CuAlMn in which CuAlMn alloy powder is dispersed in Bi
(2) Material with Cu added to Bi (3) Material with Ni added to Bi

(1) Bi-CuAlMn
CuAlMn alloys have martensitic transformation properties. The alloy phase of the metal having martensitic transformation properties takes either a martensitic phase or a parent phase based on temperature and stress. When the alloy phase of the metal is a martensite phase, the metal is extremely flexible and can be easily changed in shape based on an external force. For this reason, the stress based on an external force is relieved. Furthermore, even if the cooling and heating cycle is repeated, the shape can be changed flexibly, so that accumulation of fatigue based on stress is suppressed. In addition, when the metal alloy phase is the parent phase, the metal undergoes phase transfer to the martensite phase based on the external force and elastically deforms, so that when the external force is unloaded, the original shape is restored. Can do. For this reason, the stress applied to the metal is relieved and the accumulation of the stress is suppressed.

  Therefore, by adding a CuAlMn alloy having a martensitic transformation property to Bi which is a bulk metal, stress from an external force can be relieved and the accumulation of the stress can be controlled. As a result, the weak shear strength and brittleness characteristic of Bi are eliminated.

  Furthermore, the CuAlMn alloy has little toxicity and has little influence on the melting point (liquidus temperature or solidus temperature) of the bulk metal to be added. Further, since the CuAlMn alloy has a small electric resistance, it can be suitably used even under a situation where a current flows through the CuAlMn alloy.

The method for preparing the CuAlMn alloy powder is not particularly limited, and a known method for preparing the alloy powder can be appropriately applied. An example of the adjustment method is shown below, but is not limited thereto.
First, Cu, Al, and Mn are melted in a high-frequency melting furnace in an Ar atmosphere to prepare a CuAlMn alloy ingot as a precursor. You may add the below-mentioned additional element to an ingot as needed. Next, the obtained ingot is pulverized using a powder production technique such as an atomizing method to obtain a CuAlMn alloy powder.

  In the CuAlMn alloy, it is preferable that the Mn content is 0.01% by mass or more and 20% by mass or less, the Al content is 3% by mass or more and 13% by mass or less, and the balance is Cu. By adjusting to this composition ratio, the property of martensitic transformation appears remarkably, and it is possible to suppress the breakage of the joint formed by the solder.

Moreover, when Ag, Ni, Au, Sn, P, Zn, Co, Fe, B, Sb, and Ge are added to the CuAlMn alloy, there is an effect of improving the consistency with Bi and stabilizing the martensite phase. Therefore, an embodiment in which these additional elements are added is also preferable.
The content of the additive element in the CuAlMn alloy is preferably 0.001 to 10% by mass. When the additive element is less than 0.001% by mass, it is difficult to obtain the above effect of adding the additive element. If there are more additive elements than 10% by mass, the CuAlMn alloy cannot exhibit a martensite phase.

  It is also preferable to plate the surface of the CuAlMn alloy powder with Ni or Au. The dispersibility of the CuAlMn powder can be improved by adjusting the film thickness of the plating layer on the powder surface. The film thickness of a suitable plating layer is 0.01 μm or more and 5 μm or less, and more preferably 0.01 μm or more and 3 μm or less.

By adjusting the particle size of the CuAlMn alloy powder, the stress relaxation ability of Bi—CuAlMn can be adjusted. Specifically, the particle diameter of the CuAlMn alloy powder is preferably 0.01 μm or more and 50 μm or less, and more preferably 0.01 μm or more and 20 μm or less.
Further, the particle diameter of the CuAlMn alloy powder is 0.00005 times or more and 2.5 times or less of the particle diameter of the Bi powder (the particle diameter of the CuAlMn alloy powder / the particle diameter of the Bi powder). From the viewpoint of uniform mixing, it is preferably 0.0001 times or more and 0.5 times or less, and more preferably 0.0005 times or more and 0.25 times or less.

  The shape of the CuAlMn alloy powder may be any of a spherical shape, a rod shape, an indefinite shape, etc., but a spherical powder is preferable from the viewpoint of thermal stability and stress relaxation in the solder after sheet molding and strength improvement. .

  The content of the CuAlMn alloy in Bi—CuAlMn is preferably 0.5% by mass or more and 20% by mass or less, and more preferably 1% by mass or more and 15% by mass or less. When the content of the CuAlMn alloy is within the above range, the above effect of adding a substance having martensitic transformation properties can be obtained, and the high melting point, which is a characteristic of Bi, can be maintained. The content rate of the CuAlMn alloy in Bi-CuAlMn can be adjusted by mix | blending CuAlMn alloy powder and Bi powder by the said mass ratio.

(3) Material in which Cu is added to Bi In a solder material in which Cu is added to Bi, the brittleness of Bi is suppressed, and the mechanical strength is increased. The cause of this is not clarified, but it is considered that fine Cu is dispersed in Bi.
From the viewpoint of suppressing the brittleness of Bi, the Cu content is preferably 0.01% by mass or more, more preferably 0.1% by mass or more, and further preferably 0.4% by mass or more. It is.
On the other hand, when a large amount of Cu is added, the liquidus temperature rises. Therefore, considering the heating temperature at the time of joining with solder, the Cu content is preferably 5% by mass or less, more preferably 2% by mass or less, and further preferably 1% by mass or less.

  The Cu-containing Bi solder foil can be prepared without preparing a Cu-containing Bi ingot in advance. Specifically, Cu simple powder and Bi simple powder may be prepared, and these powders may be supplied to the powder rolling apparatus at a desired ratio.

  The particle size of Cu powder is preferably 0.00005 times or more and 2.5 times or less, and 0.0001 times or more with respect to the particle size of Bi powder (particle size of Cu powder / particle size of Bi powder). It is more preferably 0.5 times or less, and further preferably 0.0005 times or more and 0.25 times or less.

  The shape of the Cu powder may be any of a spherical shape, a rod shape, and an indefinite shape, but a spherical powder is preferred from the viewpoint of thermal stability in the solder after sheet molding and strength improvement.

(4) Material in which Ni is added to Bi In a material in which Ni is added to Bi, the brittleness of Bi is improved and the mechanical strength is increased. Although the cause of this is not clear, it is considered that the fine Bi ₃ Ni compound phase is dispersed in Bi.

From the viewpoint of keeping Bi brittleness low, the Ni content is preferably 0.01% by mass or more, more preferably 0.1% by mass or more, and even more preferably 0.4% by mass. That's it.
On the other hand, when a large amount of Ni is added, the liquidus temperature rises in the same manner as when Cu is added. Therefore, considering the heating temperature at the time of joining with solder, the Ni content is preferably 7% by mass or less, more preferably 2% by mass or less, and further preferably 1% by mass or less.

  The Ni-containing Bi solder foil can be prepared without preparing a Ni-containing Bi ingot in advance. Specifically, Ni simple powder and Bi simple powder may be prepared, and these powders may be supplied to the powder rolling apparatus at a desired ratio.

  The particle size of the Ni powder is preferably 0.00005 times or more and 2.5 times or less, and 0.0001 times or more with respect to the particle size of the Bi powder (Ni particle size / Bi powder particle size). It is more preferably 0.5 times or less, and further preferably 0.0005 times or more and 0.25 times or less.

  The shape of the Ni powder may be any of a spherical shape, a rod shape, and an indefinite shape, but a spherical powder is preferable from the viewpoints of thermal stability in the solder after sheet molding and strength improvement.

[Bi-containing solder foil]
The pure Bi solder foil 2 and the Bi-based solder foil 8 (hereinafter referred to as “Bi-containing solder foil”) obtained by the above-described manufacturing method are dense and uniform foils formed by closely adhering powders used as raw materials. . A Bi-containing solder foil having a desired composition can be obtained by changing the type of raw material supplied to the powder rolling apparatus and the blending ratio thereof.

The thickness of the Bi-containing solder foil can be set to a desired thickness by adjusting the distance and pressure between the rolling rolls 6 and 7. Specifically, the thickness of the Bi-containing solder foil can be 5 μm or more and 500 μm or less, and it is desirable that the thickness is 30 μm or more and 200 μm or less in consideration of the thickness of the joint portion in the joined body described later. 50 μm or more and 150 μm or less is more desirable. However, the thickness of the Bi-containing solder foil can be appropriately adjusted depending on the application.
Here, the thickness of the Bi-containing solder foil refers to an average thickness when any five locations are measured with a micrometer.

[Joint body of the first embodiment]
In FIG. 3, the schematic of the joined body of 1st embodiment is shown. The joined body 300 includes a first member 101 and a second member 102, and the first member 101 and the second member 102 are joined by the Bi-containing solder foil 9 described above.
Since the Bi-containing solder foil 9 is dense and uniform, unjoined portions due to voids or the like are reduced, and a dense and uniformly joined body can be obtained. Further, unlike the case where a pellet-like or lump-like solder material is applied, there is no need for external pressure and sliding called scrub while heating at the time of joining. Therefore, the work at the time of joining is also simplified.

A method for manufacturing the bonded body 300 will be described.
The Bi-containing solder foil 9 is sandwiched between the first member 101 and the second member 102 (FIG. 3A). In this laminated state, bonding is performed using a reflow method or the like to form a bonded body 300 (FIG. 4B).
The joining is preferably performed at a temperature about 30 ° C. to 60 ° C. higher than the melting point of the Bi-containing solder foil 9. The joining is performed in an inert gas or reducing gas atmosphere.

[Joint body of the second embodiment]
In FIG. 4, the schematic of the conjugate | zygote of 2nd embodiment is shown. In the joined body 310 of the second embodiment, copper (Cu) layers 111 and 112 are provided on the surfaces to be joined of the first member 101 and the second member 102, respectively. By providing the Cu layers 111 and 112, generation of unnecessary reaction products at the interface with the Bi-containing solder material 9 can be suppressed, and resistance to temperature changes is increased.

The thicknesses of the Cu layers 111 and 112 are preferably 0.1 μm or more and 10 μm or less, and more preferably 0.5 μm or more and 8 μm or less. Within this film thickness range, it is difficult for the solder material to melt and disappear during bonding, and the influence on the thermal expansion coefficient of the entire bonded body 310 is small.
The Cu layers 111 and 112 can be formed by sputtering, plating, vapor deposition, or the like.

Furthermore, a thin Au layer (not shown) may be provided on the surfaces of the Cu layers 111 and 112 in order to prevent oxidation and secure touch. Note that this thin Au layer dissolves in the solder bath at the time of joining, and hardly remains in the final power semiconductor module.
The thickness of the Au layer is preferably 0.01 μm or more and 0.5 μm or less, and more preferably 0.03 μm or more and 0.1 μm or less.

A method for manufacturing the bonded body 310 will be described.
The Cu layers 111 and 112 provided on the surfaces of the first member 101 and the second member 102 are arranged so as to face each other, and the Bi-containing solder foil 9 is sandwiched therebetween (FIG. 4A). In this stacked state, bonding is performed using a reflow method or the like to form a bonded body 310 (FIG. 4B).
The joining is preferably performed at a temperature about 30 ° C. to 60 ° C. higher than the melting point of the Bi-containing solder foil 9. The joining is performed in an inert gas or reducing gas atmosphere.

[Power semiconductor module]
In FIG. 5, the cross-sectional schematic diagram of the principal part of the power semiconductor module 10 of this embodiment is shown.
The power semiconductor module 10 of this embodiment includes a power semiconductor element 20, an insulating unit 30, and a heat sink 40. The power semiconductor element 20 and the insulating part 30 are joined by the first joining part 500. The insulating part 30 and the heat sink 40 are joined by the second joining part 600.

  The power semiconductor module 10 is used for an in-vehicle inverter or the like. Since an internal combustion engine (not shown) is provided around the power semiconductor module 10, the environment in which the power semiconductor module 10 is arranged becomes considerably high. Further, when next-generation GaN or SiC is used as the power semiconductor element, heat generated from the power semiconductor element 20 is large, and the temperature of the power semiconductor module 10 rises. In order to prevent the power semiconductor element from being destroyed by these heats, a cooling pipe (not shown) through which cooling water flows is provided, and a heat radiating plate 40 is provided between the cooling pipe and the power semiconductor element.

  In the power semiconductor module 10 of the present embodiment, at least one of the first joint 500 and the second joint 600 may be joined by the Bi-containing solder foil 9, and the first joint 500 and the second joint 600 may be joined by the Bi-containing solder foil 9.

  In the power semiconductor module 10 which is one aspect of the joined body, the Bi-containing solder foil 9 is joined to the first joint 500 or the second joint 600, so that the joint is sufficient for the operation of the power semiconductor module. Has heat resistance. Further, since the Bi-containing solder foil 9 is dense and uniform, generation of voids can be suppressed, and a dense and uniform joint can be formed. Moreover, since generation | occurrence | production of the unjoined part by a void etc. is suppressed, the increase in the thermal resistance and electrical resistance in a junction part can be suppressed. Further, unlike the case of applying a pellet-like or lump-like solder material, there is no need for external pressure and sliding called scrub while heating at the time of joining. Therefore, the work at the time of joining is also simplified.

In the power semiconductor module of the present embodiment, the first joint portion 500 may be joined first, and then the second joint portion 600 may be joined, or after the second joint portion 600 is joined first, 500 may be joined.
However, in any case, if the temperature of the second soldering is higher than the melting point of the solder material used for the first time, the portion soldered for the first time will be melted during the second soldering. It causes problems such as shifting and tilting.

  In order to avoid this problem, the solder material is selected so that the melting point of the solder material used for the first time is higher than the melting point of the solder material used for the second time. Preferably, the melting point of the solder material used for the second bonding is preferably 30 ° C. or lower than the melting point of the solder material used for the first bonding.

  In order not to destroy the power semiconductor element 20 by soldering, the upper limit of the heating temperature during soldering is 650 ° C., more preferably 450 ° C. It is preferable to use for joining.

When the Bi-containing solder foil 9 is applied to the second bonding, the solder material used for the first bonding is desirably a material having a melting point sufficiently higher than the melting point 270 ° C. of the Bi-containing solder foil 9. Thus, as the solder material used for the first bonding, an alloy (melting point main component like can be mentioned alloy material is Zn, represented by _{Zn (1-x-y)} Al x M y as an example : 382 ° C.).
In Zn _(1-x-y) alloy represented by _Al x _{M y,} the content of Al (range x) is preferably not more than 10 wt% or more 2 wt%, 3 wt% or more and 8 mass % Or less is more preferable. Metal M in the alloy represented by _{Zn (1-x-y)} Al x M y represents a metal other than zinc and aluminum, and the like Cu. In Zn _(1-x-y) alloy represented by _Al x _{M y,} the content of the metal M as (range y) is 0 to 2 wt%, preferably 0 to 1.5 mass% is there.

  It is preferable to provide a Cu layer on the surfaces to be joined of the members to be joined by the Bi-containing solder material. By providing the Cu layer, it is possible to suppress generation of unnecessary reaction products at the interface with Bi. When a reaction product is generated in a harsh cooling cycle such as a semiconductor module, a crack is generated starting from the position where the reaction product exists, and in the case of a fragile reaction product, the reaction product is cracked. Can cause cracks.

The thickness of the Cu layer provided on the surface to be joined of the Bi-containing solder foil 9 is preferably 0.1 μm or more and 10 μm or less, and more preferably 0.5 μm or more and 5 μm or less. If it is thinner than 0.1 μm, it may be dissolved into the solder material during bonding, and if it is thicker than 10 μm, it will affect the thermal expansion coefficient of the entire power semiconductor module and cause thermal stress. .
The Cu layer can be formed by sputtering, plating, vapor deposition, or the like.

The Bi-containing solder foil 9 is applied to the second joint portion 600 on the side far from the semiconductor element, and the melting point is higher than that of the Bi-containing solder foil 9 on the first joint portion 500 on the side close to the semiconductor element that generates a large amount of heat. It is preferable to apply a high solder material.
Therefore, in the power semiconductor module according to the first embodiment described below, the power semiconductor module will be described in which the Bi-containing solder foil 9 is used for the second joint portion 600.

[Power Semiconductor Module of First Embodiment]
(Power semiconductor element)
As the power semiconductor element 20, it can apply suitably according to a use without a restriction | limiting in particular, A general Si substrate etc. can also be applied.

(Insulation part)
The insulating substrate 32 in the insulating portion 30 can be applied without particular limitation as long as it can ensure insulation, but preferably has a thermal expansion coefficient of the semiconductor element so as not to cause a significant thermal stress during a cooling cycle. It has a thermal expansion coefficient comparable to the above.

Specific examples of the insulating substrate 32 include those formed of AlN, Si ₃ N ₄ , Al ₂ O _{3 and the} like. Among these, from the viewpoint of thermal conductivity and thermal expansion coefficient, Si ₃ N ₄ or AlN is preferred. Furthermore, the breaking strength of Si ₃ N ₄ is higher than the breaking strength of AlN, and the life of the insulating member itself becomes longer. Therefore, the insulating substrate 32 is more preferably Si ₃ N ₄ .

  Further, a conductive layer 34 is provided on the surface of the insulating substrate 32 in order to conduct electricity from the surface of the insulating substrate 32 on the power semiconductor element side to the semiconductor element. In order to suppress warping against temperature changes, it is preferable to provide the conductive layer 36 also on the heat radiating plate 40 side.

In the case of using AlN ceramics as the insulating substrate 32, the conductive layers 34 and 36 are preferably Al, Cu, Mo, Ni, etc. Among them, Al and Cu are preferable. When an Al layer is provided on the surface of AlN, plastic deformation occurs due to temperature changes and thermal stress can be relaxed. When a Cu layer is provided, electrical conductivity is high, so that it can be thinned and thermal stress can be relaxed. It is.
That is, when AlN ceramics are applied as the insulating substrate 32, the insulating portion 30 is preferably a Cu / AlN / Cu laminated body or an Al / AlN / Al laminated body.

On the other hand, when Si ₃ N ₄ ceramics is used as the insulating substrate 32, the conductive layers 34 and 36 are preferably Al, Cu, Mo, Ni, etc. Among them, Al and Cu are preferable, and Cu is preferable. It is more preferable. Since Cu has a high electrical conductivity, it is preferable to provide a Cu layer because it has a high electrical conductivity and can be thinned and thermal stress can be reduced. Further, the Cu / Si ₃ N ₄ / Cu laminated body is not easily broken even in the thermal cycle test, and the reliability as the insulating portion 30 is high.

  The thickness of the conductive layers 34 and 36 provided on the insulating substrate 32 is preferably 0.01 mm or more and 1 mm or less, and more preferably 0.1 mm or more and 0.8 mm or less. When the thickness of the conductive layers 34 and 36 is less than 0.01 mm, the loss and heat generation due to the current in the lateral direction cannot be ignored. When the thickness exceeds 1 mm, the thermal expansion coefficient of the entire power semiconductor module is affected. This is not preferable because it causes thermal stress.

  A method for attaching the conductive layers 34 and 36 to both surfaces of the insulating substrate 32 is not particularly limited, and a known method such as brazing can be appropriately employed.

It is preferable to provide a Cu layer (not shown) on the surface of the insulating portion 30 on the second joint portion 600 side to which the Bi-containing solder foil 9 is applied. However, when the conductive layer 36 is a Cu layer, it is not necessary to further provide a Cu layer.
As described above, the Bi-containing solder foil 9 has a very slow growth rate of products existing at the interface due to the thermal cycle at the interface with the Cu layer. Do not cause defects such as peeling.

  When a Cu layer is separately provided at the interface of the Bi-containing solder foil 9, the thickness of the Cu layer is preferably 0.1 μm or more and 10 μm or less, and more preferably 3 μm or more and 8 μm or less. If it is thinner than 0.1 μm, it may be dissolved into the solder material during bonding, and if it is thicker than 10 μm, it may affect the thermal expansion coefficient of the entire power semiconductor module.

Furthermore, a thin Au layer (not shown) may be provided on the surface of the Cu layer in order to prevent oxidation and ensure touchability. Note that this thin Au layer dissolves in the solder bath at the time of joining, and hardly remains in the final power semiconductor module.
The thickness of the Au layer is preferably about 0.01 μm to 0.5 μm, and more preferably 0.03 μm to 0.1 μm.

  A metal film (not shown) can be appropriately provided on the surface of the insulating portion 30 on the first joint portion 500 side according to the solder material of the first joint portion 500.

(Heatsink)
The heat sink 40 can be applied without particular limitation as long as it has heat dissipation properties, but it has a sufficiently high thermal conductivity and an excellent function as a heat sink and is close to the thermal expansion coefficient of the semiconductor element. It is preferable to use it.

  Specific examples of the suitable heat sink 40 include those formed of Mo, Cu—Mo alloy, Al—SiC, Cu, Al, etc. Among them, high thermal conductivity and heat close to power semiconductor elements. Mo is preferable because it has an expansion coefficient.

  When Mo is used for the heat dissipation plate, it is preferable to provide other metal layers on both sides of Mo from the viewpoint of enabling joining by soldering. Examples of such metal layers include Cu and Ni. Among these, Cu is preferable. In particular, it is preferable that the heat dissipation plate 40 is a laminate of a Cu layer 44 / Mo layer 42 / Cu layer 46 in which a Cu layer is provided on the surface of Mo from the viewpoint of adjusting the thermal conductivity and the thermal expansion coefficient. It is.

  Thus, when the heat sink 40 is a laminated body composed of the Cu layer 44 / Mo layer 42 / Cu layer 46, the thickness ratio of each layer is 1/5/1 to 1/12/1. It is preferable that it is 1/7/1 to 1/9/1. If the Mo layer is thinner than 1/5/1, it is not preferable because it has a thermal expansion coefficient far from that of the power semiconductor element. If the Mo layer is thicker than 1/12/1, the heat dissipation function as a heat sink is not sufficiently exhibited, which is not preferable.

  As a specific layer thickness, the Cu layers 44 and 46 are preferably 0.05 mm or more and 1 mm or less, and more preferably 0.2 mm or more and 0.5 mm or less. The thickness of the Mo layer 42 is preferably 1 mm or greater and 7 mm or less, and more preferably 2 mm or greater and 4 mm or less.

  In order that the laminate composed of the Cu layer 44 / Mo layer 42 / Cu layer 46 sufficiently exhibits the heat dissipation function, the total thickness is preferably 1 mm or more and 8 mm or less, and preferably 2 mm or more and 5 mm or less. Is more preferable.

It is preferable to provide a Cu layer (not shown) on the surface of the heat sink 40 on the second joint portion 600 side to which the Bi-containing solder foil 9 is applied. However, when the layer 44 is a Cu layer, it is not necessary to further provide a Cu layer.
As described above, the Bi-containing solder foil 9 has a very slow growth rate of products existing at the interface due to the thermal cycle at the interface with the Cu layer. Do not cause defects such as peeling.

  When a Cu layer is separately provided on the surface of the layer 44, the thickness of the Cu layer is preferably 0.1 μm or more and 10 μm or less, and more preferably 3 μm or more and 8 μm or less. If it is thinner than 0.1 μm, it may be dissolved into the solder material during bonding, and if it is thicker than 10 μm, it may affect the thermal expansion coefficient of the entire power semiconductor module.

Furthermore, a thin Au layer (not shown) may be provided on the surface of the Cu layer in order to prevent oxidation and ensure touchability. Note that this thin Au layer dissolves in the solder bath at the time of joining, and hardly remains in the final power semiconductor module.
The thickness of the Au layer is preferably about 0.01 μm to 0.5 μm, and more preferably 0.03 μm to 0.1 μm.

An example of the manufacturing procedure of the power semiconductor module of 1st embodiment is shown.
First, the first joint 500 is joined.
Power semiconductor element 20 and the insulating portion 30 is arranged to face the table in between the Bi content higher melting point than solder foil 9 soldering material (e.g., the above-mentioned _{Zn (1-x-y)} Al x M y Sandwiched alloy (melting point: 382 ° C.). In the laminated state, the first joining portion 500 is formed by joining using a reflow method or the like. The joining temperature is preferably 30 to 60 ° C. higher than the melting point of the solder material to be applied. The bonding is performed in an inert gas or reducing gas atmosphere. As described above, in order to reduce the oxide film formed on the surface of the coating layer, it is preferably performed in a reducing gas atmosphere.
The thickness of the formed first joint portion 500 is preferably 5 μm or more and 500 μm or less, more preferably 10 μm or more and 200 μm or less, and more preferably 50 μm or more and 150 μm or less from the viewpoint of thermal conduction and thermal stress. More preferably.

Next, the second joint 600 is joined.
In joining by the 2nd junction part 600, it arrange | positions so that the insulation part 30 and the heat sink 40 may oppose, pinch | interpose Bi containing solder foil 9 between them, and it joins using the reflow method etc. in this laminated state. The power semiconductor module 10 is obtained. The insulating unit 30 is bonded to the power semiconductor element 20 in advance by the first bonding unit 500.

Similar to the bonding by the first bonding portion 500, the bonding by the second bonding portion 600 is also performed in an inert gas or reducing gas atmosphere. The bonding temperature of the second bonding portion 600 is higher by about 30 ° C. to 60 ° C. than the melting point of the Bi-containing solder foil 9 of the second bonding portion 600 and the melting point of the solder material used for the first bonding portion 500. It is preferable to join at a temperature lower by 30 ° C. or more.
The thickness of the second joint portion 600 formed by the Bi-containing solder foil 9 is preferably 5 μm or more and 500 μm or less, more preferably 10 μm or more and 300 μm or less, from the viewpoint of thermal conduction and thermal stress, and 50 μm. More preferably, it is 150 μm or less.

Since the melting point of the solder material used for the first bonding is higher than the melting point of the Bi-containing solder foil 9 used for the second time, the portion soldered for the first time is melted and misaligned during the second soldering. There is no problem of waking up or tilting.
Further, in the first embodiment, a solder material having a higher melting point is applied to the first joint portion 500 on the side close to the semiconductor element that generates a large amount of heat, and the second joint portion 600 on the side far from the semiconductor element is applied. Applies a Bi-containing solder foil 9 having a melting point lower than that of the solder material used for the first joint portion 500, so that it becomes a power semiconductor module further excellent in heat resistance.

[Power Semiconductor Module of Second Embodiment]
In the power semiconductor module of the first embodiment, the Bi-containing solder foil 9 is applied to the second joint portion 600, and a solder material having a higher melting point than the Bi-containing solder foil 9 is applied to the first joint portion 500. In the power semiconductor module of the second embodiment, the Bi-containing solder foil 9 is applied to the first joint portion 500, and a solder material having a higher melting point than the Bi-containing solder foil 9 is applied to the second joint portion 600.
In the power semiconductor module of the second embodiment, the second joint 600 is joined first, and the first joint 500 is joined the second time. It is preferable to provide a Cu layer on the surface to be joined of the first joint portion 500 using the Bi-containing solder foil 9, that is, on the surfaces of the power semiconductor element 20 and the insulating portion 30. However, when the conductive layer 34 of the insulating unit 30 is a Cu layer, it is not necessary to further provide a Cu layer on the surface of the insulating unit 30. Since others are the same as those of the power semiconductor module of the first embodiment, description thereof is omitted.

[Power Semiconductor Module of Third Embodiment]
In the power semiconductor module of the third embodiment, the Bi-containing solder foil 9 is applied to the first joint portion 500, and a solder material having a melting point lower than that of the Bi-containing solder foil 9 is applied to the second joint portion 600. However, the solder material applied to the second bonding portion 600 has a melting point of 200 ° C. or higher in consideration of heat generation from the semiconductor element.
In the power semiconductor module of the third embodiment, the first joint 500 is first joined by the Bi-containing solder foil 9, and the second joint 600 is joined the second time. It is preferable to provide a Cu layer on the surface to be joined of the first joint portion 500 using the Bi-containing solder foil 9, that is, on the surfaces of the power semiconductor element 20 and the insulating portion 30. However, when the conductive layer 34 of the insulating unit 30 is a Cu layer, it is not necessary to further provide a Cu layer on the surface of the insulating unit 30. Since others are the same as those of the power semiconductor module of the first embodiment, description thereof is omitted.

[Power Semiconductor Module of Fourth Embodiment]
In the case where the melting point changes greatly by changing the type of alloy composition, the type of additive, and the amount of additive contained in the Bi-containing solder foil 9, Bi is added to both the first joint and the second joint. A contained solder material can be applied.
Even in this case, the melting point of the solder material used for the second bonding is desirably 30 ° C. or more lower than the melting point of the solder material used for the first bonding, and 200 ° C. or more in consideration of heat generation from the power semiconductor. It is desirable that
In consideration of heat generation from the power semiconductor element 20, it is preferable to apply a solder material having a higher melting point to the first joint, and apply a solder material having a lower melting point to the second joint.
In the power semiconductor module of the fourth embodiment, it is preferable to provide a Cu layer on the surfaces to be joined of the first joint 500 and the second joint 600. However, when the conductive layers 34 and 36 of the insulating part 30 are Cu layers, it is not necessary to further provide a Cu layer on the surface of the insulating part 30. Further, when the heat sink 40 includes the Cu layer 44, it is not necessary to further provide a Cu layer on the surface of the heat sink 40. Since others are the same as those of the power semiconductor module of the first embodiment, description thereof is omitted.

  In the power semiconductor modules of the first to fourth embodiments, the Cu layer is provided on the surface of the surface to be joined that is joined by the Bi-containing solder foil 9, but this is not always necessary and may not be provided.

  Hereinafter, the present invention will be described by way of examples, but an example of a method for producing a Bi-containing solder foil according to the present invention will be described, and the present invention is not limited to these examples.

[Example 1]
(Preparation of pure Bi powder)
The pure Bi ingot was crushed with a hammer or the like, and then crushed with a rough machine to prepare pure Bi powder. When the shape of the obtained pure Bi powder was observed with an electron microscope, it was a rod shape having a long side of 150 to 50 μm and a short side of 80 to 20 μm. Moreover, it was 70 micrometers when the average particle diameter of pure Bi powder was measured by the above-mentioned method.

(Preparation of CuAlMn alloy powder)
Cu 82 mass%, Al 9 mass%, and Mn 9 mass% were melted in a high-frequency melting furnace in an Ar atmosphere to prepare a CuAlMn alloy ingot having a martensite phase. The obtained ingot was dissolved in an argon gas atmosphere and pulverized by a gas atomization method with an injection pressure of 5 MPa to obtain a CuAlMn alloy powder. This CuAlMn alloy powder was classified with sieves of 250 μm, 150 μm, 53 μm, and 20 μm, and powders of 53 μm or more were removed.
When the average particle size of the obtained CuAlMn alloy powder was measured by the above-mentioned method, the average particle size was 36 μm, and the particle size distribution width was measured by the laser diffraction scattering method to be 2.0. Moreover, when the shape of the obtained CuAlMn alloy powder was observed with the electron microscope, the shape was spherical.

The classified CuAlMn alloy powder was plated with Ni using a dropping method.
When the film thickness of the Ni plating layer was measured by cross-sectional observation, it was 5 μm even at a thick part. The average film thickness of the Ni plating layer when measuring 10 CuAlMn alloy powders was 0.8 μm.

(Powder rolling)
100 g of the prepared pure Bi powder and 10 g of Ni-plated CuAlMn alloy powder were put into a container and stirred sufficiently, and then the powder was loaded into a powder rolling machine manufactured by Ono Roll Seisakusho as shown in FIG. The rolling conditions are as follows.
・ Temperature: 20 ℃
・ Roll diameter: 15cm
-Distance between rolls: 240 μm
・ Roll rotation speed: 8rpm

(Bi-containing solder foil)
FIG. 6 shows an appearance photograph of the obtained Bi-containing solder foil, and FIG. 7 shows an electron micrograph of its cross-sectional structure. The obtained Bi-containing solder foil had a width of 2 cm, a length of 10 cm or more, and a thickness of 250 μm. In addition, it has confirmed that a longer solder foil can be produced by changing rolling conditions.
The cross-sectional structure photograph shown in FIG. 7 confirmed that the CuAlMn alloy powder was uniformly dispersed in Bi.

[Example 2]
The power semiconductor element and the insulating substrate were joined with the Bi-containing solder foil produced in Example 1, and the evaluation test body-1 having the configuration shown in FIG. 8 was produced.
Cu layers 34 and 36 having a thickness of 150 μm were attached to both surfaces of a ceramic plate of Si ₃ N ₄ by brazing to produce a laminate of Cu layer 34 / Si ₃ N ₄ layer 32 / Cu layer 36.
Moreover, the power semiconductor element 20 using GaN was prepared, and the Cu layer 22 was formed on the outermost surface by sputtering. When the thickness of the Cu layer 22 was confirmed with an electron microscope, it was about 3 μm.

The Cu layer 22 of the prepared power semiconductor element 20 and the Cu layer 34 of the insulating portion 30 are arranged so as to face each other, and positioning is performed with the Bi-containing solder foil produced in Example 1 sandwiched therebetween. Was set in a carbon jig and placed in an electric furnace.
Nitrogen gas containing 5% by volume of hydrogen gas was always introduced into the core tube of the electric furnace, and the inside was made a reducing atmosphere. In this state, the temperature was raised from room temperature to 320 ° C. over 30 minutes, and then kept at 320 ° C. for 5 minutes.

<Observation of joint surface>
Evaluation test body-1 first observed the entire joint surface in a nondestructive manner by ultrasonic flaw detection, then cut evaluation test body-1 to polish the cross section, and observed the cross section with an electron microscope.
As a result, no void was confirmed on the joint surface of the evaluation test body-1, and it was found that the joint was dense and uniform.

[Comparative Example 1]
100 g of pure Bi powder prepared in Example 1 and 10 g of Ni-plated CuAlMn alloy powder were dissolved in an argon gas atmosphere and cast into a mold to prepare an ingot. After polishing the surface, it was cut out to 1.0 cm × 1.0 cm × 1.0 cm and attempted to be rolled to a target thickness of 250 μm by cold rolling. However, when the thickness reached 0.8 mm, the sample was cracked and a sheet could not be formed.

[Example 3]
In the same manner as in Example 1, a pure Bi ingot was crushed with a hammer or the like, and then pulverized with a rough machine to prepare pure Bi particles. When the pulverization time of the rough machine was halved and the shape of the obtained pure Bi particles was observed with an electron microscope, it was a rod having a long side of 250 to 150 μm and a short side of 80 to 20 μm. Moreover, it was 170 micrometers when the average particle diameter of pure Bi particle | grains was measured by the above-mentioned method.

  When this pure Bi particle was rolled by the same method as in Example 1, a Bi-containing solder foil was obtained. When this Bi-containing solder foil was observed with a cross-sectional structure photograph, it was confirmed that the CuAlMn alloy powder was uniformly dispersed in Bi.

[Example 4]
A pure Cu plate was dissolved in an argon gas atmosphere and pulverized by a gas atomization method with an injection pressure of 5 MPa to obtain Cu powder. This Cu powder was classified with sieves of 250 μm, 150 μm, 53 μm, and 20 μm, and powders of 53 μm or more were removed.
When the average particle size of the obtained Cu powder was measured by the above-mentioned method, the average particle size was 25 μm, and the particle size distribution width was measured by a laser diffraction scattering method, which was 1.8. Moreover, when the shape of the obtained Cu powder was observed with the electron microscope, the shape was spherical.

In the same manner as in Example 1, except that the CuAlMn alloy powder was replaced with the Cu powder and rolled, a Bi-containing solder foil could be obtained.
When the obtained Cu-containing Bi-containing solder foil was observed with a cross-sectional structure photograph, it was confirmed that the Cu powder was uniformly dispersed in Bi.

[Example 5]
Rolling was performed in the same manner as in Example 1, except that the roll rotation speed was changed from 8 rpm to 15 rpm, and the distance between rolls was changed from 240 μm to 25 μm.
The obtained Bi-containing solder foil had a width of 5 cm or more and a length of 15 cm or more and a thickness of 30 μm. It has been confirmed that a longer solder foil can be produced by changing the rolling conditions.

It is a schematic diagram of the manufacturing apparatus (powder rolling machine) used for the manufacturing method of Bi containing solder foil of 1st embodiment. It is a figure explaining the manufacturing method of Bi containing solder foil of a second embodiment. It is the schematic of the conjugate | zygote of 1st embodiment. It is the schematic of the conjugate | zygote of 2nd embodiment. It is a cross-sectional schematic diagram of the principal part of the power semiconductor module 10 of this embodiment. 2 is an appearance photograph of a Bi-containing solder foil obtained in Example 1. FIG. 2 is an electron micrograph of a cross-sectional structure of a Bi-containing solder foil obtained in Example 1. FIG. 4 is a schematic configuration diagram of an evaluation test body-1 of Example 2. FIG.

Explanation of symbols

2 Pure Bi solder foil 4 Bi powder 6, 7 Roll roll 8 Bi-based solder foil 9 Bi-containing solder foil 10 Power semiconductor module 20 Power semiconductor element 22, 38 Copper (Cu) layer 30 Insulating part 32 Insulating substrate 34, 36 Conductive layer 40 Heat sink 42 Molybdenum (Mo) layer 44, 46 Copper (Cu) layer 101 First member 102 Second member 111, 112 Copper (Cu) layer 300, 301 Joint 500 First joint 600 Second joint

Claims (13)

Bi powder and CuAlMn alloy powder are mixed, and a powder of solder material containing bismuth (Bi) containing 30 mass% or more of Bi powder is formed into a sheet by a powder rolling method. Method. A method for producing a Bi-containing solder foil comprising mixing a Bi powder and a Cu powder, and forming a powder of a solder material containing bismuth (Bi) containing 30% by mass or more of the Bi powder by a powder rolling method . 3. The method for producing a Bi-containing solder foil according to claim 1, wherein the Bi powder has an average particle size of 5 μm to 200 μm.   The ratio of the long side to the short side of each powder particle in the Bi powder is in the range of 1.2 or more and 3.0 or less, according to any one of claims 1 to 3. A method for producing a Bi-containing solder foil. The rolling by powder rolling method, Bi-containing solder foil method according to any one of claims 1 to 4, characterized in that it is carried out at -20 ° C. or higher 269 ° C. or less. A Bi-containing solder foil comprising a powder of a solder material containing 30% by mass or more of Bi powder and further containing CuAlMn alloy powder or bismuth (Bi) containing Cu powder by a powder rolling method. The Bi-containing solder foil according to claim 6 , wherein an average thickness is 30 μm or more and 250 μm or less. A first member and a second member;
Assembly between said first and second members, characterized by comprising joined at Bi-containing solder foil according to claim 6 or claim 7. A power semiconductor element, an insulating substrate, and a heat sink;
Bi content of Claim 6 or Claim 7 in which at least one of the 1st junction part between the said power semiconductor element and the said insulation board | substrate and the 2nd junction part between the said insulation board | substrate and the said heat sink is contained. A power semiconductor module characterized by being joined by solder foil. The power semiconductor module according to claim 9 , wherein the power semiconductor element is made of GaN or SiC. Wherein the insulating substrate is a Si ₃ N ₄ layers, power of claim 9 or claim 10, characterized in that it comprises a conductive layer formed of copper (Cu) on both the surface of the Si ₃ N ₄ layer semiconductor module. The heat radiating plate, any one of claims 9 to 11 which is a laminate of Cu layer / Mo layer / Cu layer having a double-sided copper (Cu) layer of molybdenum (Mo) layer Power semiconductor module as described in 2. 13. The power semiconductor module according to claim 12 , wherein a ratio of a thickness of Cu layer / Mo layer / Cu layer in the heat radiating plate is 1/5/1 to 1/12/1.