JP5584909B2 - Connection structure - Google Patents

Description

The present invention relates is that connection structure used for connecting the electronic device.

The trend toward miniaturization, weight reduction, and high functionality of electronic devices typified by mobile phones is remarkable, and following this trend, high-density packaging technology continues to make rapid progress.
Various mounting techniques have been developed in order to incorporate components in a substrate, package a plurality of LSIs in one package, and effectively use a limited volume. On the other hand, as the density increases, the solder joints of components built into the board or package are more frequently subjected to heat treatment in the subsequent process, and the solder re-use that occurs in the gap between the component and the sealing resin occurs. The short problem due to melting has become apparent.
Therefore, there is a demand for the development of a lead-free solder material that does not melt even if it is subjected to multiple heat treatments in the subsequent process for connecting components incorporated in the substrate or package.

The present inventors have proposed a lead-free solder material that can be melt-bonded under reflow heat treatment conditions for lead-free solder and that does not melt under the same heat-treatment conditions after bonding (see Patent Document 1 below).
The reflow heat treatment conditions for lead-free solder are typical Sn-3.0Ag-0.5Cu (melting point: 217 ° C), and are general reflow heat treatment conditions for solder connection, and have a peak temperature of 240-260 ° C. It is a range.
The conductive filler of the solder material is composed of a mixture of the first metal particles mainly composed of Cu and the second metal particles that melt in the reflow heat treatment, and forms a new stable alloy phase in the reflow heat treatment. Even in the reflow heat treatment again, it has a feature that it does not melt.
In the solder material, it is necessary to increase the contact area from the viewpoint of promoting the thermal diffusion reaction between the first metal particles and the melted second metal particles in the reflow heat treatment, and fine particles having an average particle diameter of 6 μm or less are formed. It is used.

On the other hand, a solder material using a mixture of Cu particles and Sn particles having an average particle size of 30 μm as a conductive filler has been proposed (see Patent Document 2 below).
The solder material is characterized in that it is connected by a heat treatment to a CuSn compound containing Cu ₆ Sn ₅ and a connection part having Cu particles, and the Cu particles are connected by the CuSn compound.
However, in the solder material, Cu ₆ Sn ₅ formed at the Cu interface grows as a coarse crystallized substance inside the connection, so that mechanical properties such as bonding strength are likely to deteriorate, and there is a problem in connection reliability. there were.
In addition, Cu particles tend to oxidize and agglomerate and agglomerate more strongly when moisture is absorbed. Therefore, there is a problem in storage stability.

International Publication No. 2006/109573 Pamphlet Japanese Patent No. 3414388

  The present invention has been made in view of the above problems, and the problem to be solved by the present invention is that it can be melt-bonded under reflow heat treatment conditions of lead-free solder, and after the bonding, a plurality of heat treatments are performed in a subsequent process. It is to provide a lead-free solder connection structure having excellent connection reliability that does not melt even when it is received.

As a result of intensive studies to solve the above-mentioned problems and repeated experiments, the present inventors have reached the present invention .
[1] A metal filler composed of 100 parts by mass of Cu alloy particles and 30 to 550 parts by mass of Sn particles or Sn alloy particles, wherein the Cu alloy particles are composed of Cu and In, and Ag, Sn, and Bi. The metal filler is characterized in that it contains one or more metals selected from the group consisting of Cu alloy particles and the Cu alloy particles have an average particle size of 10 to 30 μm.

  [2] The metal according to [1], wherein the Cu alloy particles contain 5 to 15 mass% Ag, 2 to 8 mass% Bi, 49 to 81 mass% Cu, 2 to 8 mass% In, and 10 to 20 mass% Sn. Filler.

  [3] The Cu alloy particles have at least one metastable alloy phase observed as an exothermic peak in differential scanning calorimetry (DSC) at 230 to 300 ° C., and a melting point observed as an endothermic peak at 480 to 530 ° C. The metal filler according to [1] or [2], which has at least one.

  [4] Any of the above [1] to [3], wherein the Cu alloy particles include crystal grains of a Cu—Sn alloy phase or a Cu—In alloy phase, and Ag or Bi is present at an interface between these alloy phases. Metal filler according to crab.

  [5] A solder paste containing the metal filler according to any one of [1] to [4] and a flux component.

  [6] A connection structure obtained by reflow heat treatment of the solder paste according to [5].

That is, the present invention is as follows:
[7] A connection structure having particles containing Cu as a component in a conductive region which is a lead-free solder, wherein an alloy phase containing Cu-Sn-In is present at the interface of at least the particles containing Cu, An average particle diameter of particles containing Cu is 2 to 28 μm, and a peak position of a particle size distribution is 5 to 25 μm.

  The connection structure obtained by reflow heat treatment of the metal filler of the present invention and the solder paste containing the same has few voids, and the solder connection portion is melted even when subjected to reflow heat treatment of lead-free solder multiple times in the subsequent process. Therefore, the connection reliability is excellent.

2 is a DSC chart obtained by differential scanning calorimetry of Cu alloy particles granulated and classified in Example 1. FIG. 2 is a cross-sectional TEM (transmission electron microscope) image photograph of Cu alloy particles granulated and classified in Example 1. FIG. 2 is a reflected electron image photograph of a cross-section of a connection part produced in Example 1.

The metal filler of this embodiment is a metal filler composed of 100 parts by mass of Cu alloy particles and 30 to 550 parts by mass of Sn particles or Sn alloy particles, and the Cu alloy particles are Cu and In, and Ag, Sn, And one or more metals selected from the group consisting of Bi, and the Cu alloy particles have an average particle size of 10 to 30 μm.

When the preferred composition of the metal filler is exemplified, the mixing ratio of Sn particles or Sn alloy particles is 550 parts by mass or less from the viewpoint of heat resistance, while the lower limit is 30 from the viewpoint of improving the initial bonding state. More than part by mass.
The component ratio of the Cu alloy particles includes Cu and In and one or more selected from the group consisting of Ag, Sn, and Bi from the viewpoint of alloying by thermal diffusion with Sn or Sn alloy particles. Contains metal. Specifically, the Cu alloy particles preferably contain 5 to 15% by mass of Ag, 2 to 8% by mass of Bi, 49 to 81% by mass of Cu, 2 to 8% by mass of In, and 10 to 20% by mass of Sn.

  The particle size of the Cu alloy particles is in the range of 10 to 30 μm as an average particle size. If the average particle size is 10 μm or more, the specific surface area of the particles will be small, so there will be little reaction with the flux, so the life of the paste will be extended, and in reflow heat treatment, it will occur due to reduction by flux (particle oxide film removal) Since less gas is generated, voids are less likely to occur inside the connection, while the upper limit is 30 μm or less from the viewpoint of paste characteristics. When the particle size is increased, the gap between the particles is increased, so that the adhesive force is easily impaired.

  As the particle size of the Sn particles or the Sn alloy particles, the average particle size is preferably in the range of 2 to 30 μm, more preferably 5 to 30 μm, from the viewpoint of the reactivity with the flux and the paste characteristics, like the Cu alloy particles. It is a range.

The particle size distribution of Cu alloy particles and Sn particles or Sn alloy particles can be determined according to the paste application. For example, in screen printing applications, it is preferable to make the particle size distribution broader with an emphasis on plate slippage, and in dispensing applications, it is preferable to focus on ejection fluidity and make the particle size distribution sharper.
The Cu alloy particles have at least one metastable alloy phase observed as an exothermic peak in differential scanning calorimetry (DSC) at 230 to 300 ° C. and a melting point observed as an endothermic peak at least 480 to 530 ° C. It is preferable to have one. Heat generation in differential scanning calorimetry (DSC) is detection of latent heat generated when a new alloy phase is formed, and indicates that a metastable alloy phase is present in the alloy particles.

As a method for producing alloy particles having a metastable alloy phase, a rapid solidification method is preferable. Examples of the method for producing fine powder by the rapid solidification method include a water spray method, a gas spray method, and a centrifugal spray method. From the viewpoint of suppressing the oxygen content of particles, the gas spray method and the centrifugal spray method are more preferable. preferable.
In the gas spraying method, inert gas such as nitrogen gas, argon gas, helium gas, etc. can usually be used, but helium gas with a low specific gravity is used to increase the linear velocity during gas spraying and increase the cooling rate. Is preferably used. The cooling rate is preferably in the range of 500 to 5000 ° C./second. In the centrifugal spray method, from the viewpoint of forming a uniform molten film on the upper surface of the rotating disk, the material is preferably sialon, and the disk rotation speed is preferably in the range of 60,000 to 120,000 rpm.

  The Cu alloy particles preferably include crystal grains of a Cu—Sn alloy phase or a Cu—In alloy phase, and Ag or Bi is preferably present at the interface of the alloy phase. By having a metastable alloy phase with high reactivity such as Cu—Sn alloy phase and Cu—In alloy phase inside, it is possible to rapidly alloy with molten Sn particles or Sn alloy particles in reflow heat treatment. it can.

  The solder paste of the present invention preferably contains a metal filler and a flux component, and the content of the metal filler is preferably in the range of 84 to 94% by mass with respect to 100% by mass of the solder paste from the viewpoint of paste characteristics. The content rate of a metal filler can be defined according to a paste use. For example, the content of the metal filler is preferably in the range of 87 to 91% by mass, more preferably in the range of 88 to 90% by mass, with emphasis on plate slippage in screen printing applications, and in the dispensing application, In consideration of the discharge fluidity, the range of 85 to 89% by mass is preferable, and the range of 86 to 88% by mass is more preferable.

  The flux component preferably contains a modified rosin, a solvent, an activator, and a thixotropic agent. The flux is optimal for the surface treatment of the metal filler, removes the oxide film of the metal filler during the reflow heat treatment, and promotes the melting of the metal and alloying by thermal diffusion. A known material can be used as the flux.

In the solder paste, when a mounting component electrode such as an electronic device is connected to a substrate electrode, if a thermal history equal to or higher than the melting point of the Sn particles or Sn alloy particles is given under reflow heat treatment conditions of lead-free solder, the Sn The particles or the Sn alloy particles are melted, and the mounted component electrode and the substrate electrode are joined via the Cu alloy particles. As a result, the thermal diffusion reaction between the metals proceeds at an accelerated rate, and a new stable alloy phase having a melting point higher than the melting point of the Sn particles or Sn alloy particles is formed. To form a connection structure.
The melting point of this new stable alloy phase is higher than the reflow heat treatment temperature of lead-free solder, and since it does not melt even when subjected to a plurality of heat treatments in the subsequent process, it is possible to suppress a short circuit due to solder remelting.

The new stable alloy phase is preferably a Cu—Sn system from the viewpoint of connection reliability, and more preferably a Cu—Sn—Ni system or a Cu—Sn—In system.
In the Cu—Sn binary system, since Cu ₆ Sn ₅ grows as a coarse crystallized substance inside the connection, there is a problem that mechanical properties such as bonding strength are easily deteriorated. However, if a small amount of Ni, In, or the like is present in the alloy phase, the crystal grains are refined, so that the connection reliability is improved. The component ratio of Ni, In, etc. in the alloy phase is preferably 5% by mass or less, more preferably 3% by mass or less from the viewpoint of forming a stable alloy phase. Moreover, in order to express the connection reliability improvement effect, the component ratio is preferably 0.01% or more.

  In addition, as a connection method using a solder paste, after applying the paste to the substrate electrode, placing the mounting component and connecting by reflow heat treatment, applying the paste to the mounting component electrode or the substrate electrode, forming the bump by reflow heat treatment, For example, the component and the substrate may be combined and connected again by reflow heat treatment.

  Further, the connection structure of the present invention is a connection structure having particles containing Cu as a component in a conductive region which is a lead-free solder, wherein the alloy phase containing Cu-Sn-In contains at least the Cu. The average particle size of the particles containing Cu is 2 to 28 μm, and the peak position of the particle size distribution is 5 to 25 μm.

The conductive region that is lead-free solder is not particularly limited as long as it is a conductive region. For example, in a reflow heat treatment of lead-free solder, it is a metal portion formed by fusion bonding of a metal filler.
The Cu-containing particles are the portions that appear darkest when a cross-section of the connection structure is taken out and a reflected electron image is taken at an acceleration voltage of 10 kV and a magnification of 2000. The average particle size is an average value of particle size data of 300 or more particles, and the peak position of the particle size distribution is a peak value of the same data.

Hereinafter, the present invention will be specifically described by way of examples.
[Example 1]
(1) Production of Cu alloy particles Cu 6.5 kg (purity 99 mass% or more), Sn 1.5 kg (purity 99 mass% or more), Ag 1.0 kg (purity 99 mass% or more), Bi 0.5 kg (purity 99 mass% or more) ), And 0.5 kg of In (purity 99% by mass or more) were placed in a graphite crucible and heated and melted to 1400 ° C. with a high-frequency induction heating apparatus in a helium atmosphere of 99% by volume or more.
Next, after this molten metal is introduced from the tip of the crucible into a spray tank in a helium gas atmosphere, helium gas (purity 99 vol% or more, oxygen concentration 0.1 vol%) is supplied from a gas nozzle provided near the crucible tip. And pressure was reduced to 2.5 MPa), and atomization was performed to prepare Cu alloy particles. The cooling rate at this time was 2600 ° C./second.
When the obtained Cu alloy particles were observed with a scanning electron microscope (Hitachi, Ltd .: S-3400N), they were spherical.

The Cu alloy particles were classified using an airflow classifier (Nisshin Engineering: TC-15N) at a setting of 20 μm, the large particles were collected, and then classified again at a setting of 30 μm, and the small particles were collected. The collected alloy particles were measured with a laser diffraction particle size distribution analyzer (HELOS & RODOS), and the average particle size was 15.1 μm.
Next, Cu alloy particles were measured with a differential scanning calorimeter (Shimadzu Corporation: DSC-50) in a temperature range of 40 to 580 ° C. under a nitrogen atmosphere under a temperature rising rate of 10 ° C./min. The DSC chart obtained as a result is shown in FIG. As shown in FIG. 1, endothermic peaks were detected at 502 ° C. and 521 ° C., and the presence of a plurality of alloy phases could be confirmed from a plurality of melting points. Further, an exothermic peak was detected at 258 ° C. and 282 ° C., and the presence of a metastable alloy phase was confirmed.

  Next, the alloy particles were embedded in a resin and then cut with a microtome to observe the internal structure. FIG. 2 shows a cross-sectional TEM (transmission electron microscope) image obtained as a result. As a result of analyzing the alloy structure by TEM-EDX (characteristic X-ray analyzer), the alloy particles include Cu—Sn alloy phase or Cu—In alloy phase crystal grains, and Ag or Bi at the interface of the alloy phase. Was confirmed to exist.

(2) Manufacture of Sn particles 10.0 kg of Sn (purity 99% by mass or more) was put in a graphite crucible, and heated and melted to 1400 ° C. with a high-frequency induction heating device in a helium atmosphere of 99% by volume or more.
Next, after this molten metal is introduced from the tip of the crucible into a spray tank in a helium gas atmosphere, helium gas (purity 99 vol% or more, oxygen concentration 0.1 vol%) is supplied from a gas nozzle provided near the crucible tip. Less than the pressure of 2.5 MPa) was atomized to produce Sn particles. The cooling rate at this time was 2600 ° C./second.
When the obtained Sn particles were observed with a scanning electron microscope (Hitachi, Ltd .: S-3400N), they were spherical.

The Sn particles were classified using an airflow classifier (Nisshin Engineering: TC-15N) at a setting of 5 μm, and after collecting the large particles, they were classified again at a setting of 40 μm, and the small particles were collected. The collected Sn particles were measured with a laser diffraction particle size distribution analyzer (HELOS & RODOS), and the average particle size was 6.9 μm.
Next, Sn particles were measured with a differential scanning calorimeter (Shimadzu Corporation: DSC-50) in a temperature range of 40 to 580 ° C. under a nitrogen atmosphere under a temperature rising rate of 10 ° C./min. As a result, an endothermic peak was detected at 242 ° C., and it was confirmed that it had a melting point of 232 ° C. (melting start temperature: solidus temperature). A characteristic exothermic peak was not detected.

(3) Production of solder paste The Cu alloy particles and Sn particles were mixed at a weight ratio of 100: 82 to obtain a metal filler. Next, 89.9% by mass of a metal filler and 10.1% by mass of a rosin flux are mixed, and solder paste is sequentially applied to a solder softener (Malcom: SPS-1) and a defoaming kneader (Matsuo Sangyo: SNB-350). Was made. The solder paste thus obtained was measured with a spiral viscometer (Malcom: PCU-205). The viscosity was 194 Pa · s and the thixo index was 0.46.

(4) Measurement of bonding strength Next, a solder paste is printed on a Cu substrate having a size of 25 mm × 25 mm and a thickness of 0.25 mm, and a Cu chip having a size of 2 mm × 2 mm and a thickness of 0.5 mm is mounted, and then in a nitrogen atmosphere. A sample was prepared by reflow heat treatment at a peak temperature of 250 ° C. A reflow simulator (Malcom: SRS-1C) was used as the heat treatment apparatus. The temperature profile was raised from the start of heat treatment (room temperature) to 140 ° C. at 1.5 ° C./second, gradually increased from 140 ° C. to 170 ° C. over 110 seconds, and then from 170 ° C. to 250 ° C. The temperature was raised at a rate of 0.0 ° C./second, and the conditions were maintained at a peak temperature of 250 ° C. for 15 seconds. For the printing pattern formation, a screen printer (Microtech: MT-320TV) was used, the plate was made of metal, and the squeegee was made of urethane. The mask opening size is 2 mm × 3.5 mm, and the thickness is 0.1 mm. The printing conditions were a speed of 50 mm / second, a printing pressure of 0.1 MPa, a squeegee pressure of 0.2 MPa, a back pressure of 0.1 MPa, an attack angle of 20 °, a clearance of 0 mm, and a printing frequency of once.

  Next, the chip bond strength in the shearing direction of the prepared sample was measured at a normal speed (25 ° C.) with a push-pull gauge at a pressing speed of 10 mm / min, and converted to a unit area of 9.6 MPa. Furthermore, after heating the said preparation sample to 260 degreeC on a hotplate and hold | maintaining for 1 minute, when the chip | tip joining strength of a shear direction was measured by the same method as the above, it was 2.3 MPa, and joining strength is maintained also at 260 degreeC. The heat resistance which can be confirmed was confirmed. The results are shown in Table 1 below.

(5) Confirmation of Cu—Sn—In Alloy Phase Next, the prepared sample was resin-embedded (buhler: eposin), and then polished to obtain a cross section of the Cu chip center portion, and the connection portion was observed. The reflected electron image obtained as a result is shown in FIG. In FIG. 3, when the average particle diameter of the darkest portion (particles containing Cu as a component) of the reflected electron image was measured, it was 12.4 μm, and the peak value of the particle size distribution was 11.1 μm.
Next, the TEM-EDX (characteristic X-ray analyzer) was used to measure the alloy phase at the interface around the particles containing Cu in the connecting portion, and it was confirmed that a Cu-Sn-In alloy phase having 1.3 mass% of In was present. did.

[Examples 2 to 7]
Using the mixture in which the mixing ratio of Cu alloy particles and Sn particles described in Example 1 is changed as a metal filler, after pasting and reflow heat treatment in the same manner as in Example 1, chip bonding strength, Cu of the connection structure In Table 1 below, the average particle diameters of the particles containing are measured as Examples 2 to 7.
In all of Examples 2 to 7, it was confirmed that a Cu—Sn—In alloy phase was present at the peripheral interface of the particles containing Cu in the connection structure.

[Comparative Example 1]
Table 1 below shows the results of a typical lead-free solder Sn-3.0Ag-0.5Cu as Comparative Example 1.
As apparent from the results in Table 1, in the state heated to 260 ° C., in Comparative Example 1, the solder connection portion melts, whereas in Examples 1 to 7, there is a bonding strength of 0.2 MPa or more. It can be seen that there is sufficient heat resistance to maintain the connection state.

[Example 8]
The Cu alloy particles before classification produced by gas atomization in Example 1 were classified using an airflow classifier (Nisshin Engineering: TC-15N) at a setting of 30 μm, and after collecting the large particle side, classification was performed again at a setting of 40 μm. Using the Cu alloy particles having an average particle diameter of 26.4 μm obtained by collecting the small particle side, mixing with Sn particles, pasting, and reflow heat treatment in the same manner as in Example 1, followed by chip bonding strength and connection structure The average particle size of the particles containing Cu in the body was measured. As a result, the shear strength at normal temperature was 11.8 MPa, the shear strength at 260 ° C. was 2.1 MPa, and the average particle size of the particles containing Cu in the connection structure was 23.5 μm. there were.
It was also confirmed that a Cu—Sn—In alloy phase was present at the interface around the particles containing Cu in the connection structure.

[Example 9]
The solder paste produced in Example 1 was printed on the Cu electrode of a printed circuit board made of a high heat-resistant epoxy resin glass cloth, and a 0603 size multilayer ceramic chip capacitor (hereinafter also referred to as 0603C) and a 1005 size multilayer ceramic chip capacitor ( Hereinafter, it was also referred to as “1005C”), and a sample was prepared by reflow heat treatment at a peak temperature of 250 ° C. by the heat treatment method.
Next, the fabricated sample was heated to 80 ° C. on a hot plate, applied with an underfill to the height of the mounted component, placed in an oven, and cured at 170 ° C. for 1 hour. Next, the mold resin was applied to the upper part and the periphery of the mounted component, placed in an oven, cured at 150 ° C. for 2 hours, and then baked at 125 ° C. for 24 hours.

Next, after absorbing moisture at 60 ° C. and 60% RH for 40 hours, reflow heat treatment at a peak temperature of 260 ° C. was repeated 10 times in a nitrogen atmosphere. A reflow simulator (Malcom: SRS-1C) was used as the heat treatment apparatus. The temperature profile was as follows: from heat treatment start (room temperature) to 150 ° C. at a rate of 1.5 ° C./second, gradually increasing from 150 ° C. to 210 ° C. over 100 seconds, then from 210 ° C. to 260 ° C. The temperature was raised at a rate of 0.0 ° C./second, and the conditions were maintained at a peak temperature of 260 ° C. for 15 seconds.
Next, the prepared sample was immersed in a resin solubilizer Dynasolv (Dynaloy: 711), and ultrasonic waves were applied to remove the resin around the mounted component.
Next, the reflow heat treatment was used to observe whether or not the solder melted and the traces of the gap between the part and the sealing resin were wet and spread, but no trace was observed, and heat resistance that did not flow even at 260 ° C. was confirmed. The results are shown in Table 2 below.

[Examples 10 and 11]
Table 2 below shows the results of evaluating the solder paste produced in Example 4 as Example 10 by the same method as in Example 9.
Table 2 below shows the results of evaluating the solder paste produced in Example 6 by the same method as in Example 9 as Example 11.

[Comparative Example 2]
Table 2 below shows the results of a typical lead-free solder Sn-3.0Ag-0.5Cu evaluated as a comparative example 2 in the same manner as in Example 9.
As is clear from the results in Table 2, in Comparative Example 2, there was a very high probability that traces of solder melting and spreading the gap between the component and the sealing resin were confirmed, and the component was likely to cause a short circuit. Turned out to be. Occurrence was more pronounced with small 0603C parts.

[Example 12]
The solder paste produced in Example 1 was printed on the Cu electrode of the printed circuit board, and 0603 size chip resistor (hereinafter also referred to as 0603R), 0603C, and 1005 size chip resistor (hereinafter also referred to as 1005R). , And 1005C were mounted, and then a sample was prepared by reflow heat treatment at a peak temperature of 250 ° C. by the heat treatment method.

Next, the prepared sample was resin-embedded (Buhler: eposin) and then polished to obtain a cross section of the center portion in the longitudinal direction of the component.
In addition, when the average particle diameter of the particle | grains containing Cu of a connection structure was calculated | required for the said preparation sample by the method similar to Example 1, it was 8.3 micrometers and the peak value of the particle size distribution was 7.4 micrometers.
Next, an image of the solder connection portion was taken, and the ratio of voids contained in the solder connection portion was calculated using image processing software (Mitani Corporation: Win Roof). When the above work was performed at 10 points for each part and the average value was calculated, the void ratio was 3.6% for 0603R, 6.0% for 0603C, 6.5% for 1005R, 6.4% for 1005C, All were 10% or less, and it was confirmed that there were very few voids. The results are shown in Table 3 below.

[Examples 13 and 14]
Table 3 below shows the results of evaluating the solder paste prepared in Example 4 by the same method as in Example 12 as Example 13.
Table 3 below shows the results of evaluating the solder paste prepared in Example 6 by the same method as in Example 12 as Example 14.

[Comparative Examples 3 to 6]
In Table 3 below, as Comparative Example 3, Cu alloy particles before classification produced by gas atomization in Example 1 were classified using an airflow classifier (Nisshin Engineering: TC-15N) at a setting of 1.6 μm, After collecting the large particle side, classification is performed again at a setting of 10 μm, and Cu alloy particles having an average particle diameter of 2.7 μm obtained by collecting the small particle side are mixed with Sn particles in the same manner as in Example 1 and pasted. The results of evaluation in the same manner as in Example 12 are shown.

  Moreover, as Comparative Example 4, the Cu alloy particles before classification produced by gas atomization in Example 1 were classified using an airflow classifier (Nisshin Engineering: TC-15N) at a setting of 40 μm, and the small particle side was recovered. The Cu alloy particles having an average particle diameter of 4.4 μm obtained and the Sn particles before classification produced by gas atomization were classified using an airflow classifier (Nisshin Engineering: TC-15N) at a setting of 40 μm. The results of mixing and pasting Sn particles with an average particle diameter of 5.7 μm obtained by collecting the particle side at the same weight ratio as in Example 1 and evaluating in the same manner as in Example 12 are shown.

Further, as Comparative Example 5, the results are shown in which the Cu alloy particles and Sn particles classified in Comparative Example 4 were mixed and pasted at the same weight ratio as in Example 4, and evaluated in the same manner as in Example 12.
Further, as Comparative Example 6, Cu particles having an average particle diameter of 15.4 μm (Fukuda Metal Foil Powder Industry: Cu-HWQ 15 μm) were used instead of Cu alloy particles, and mixed with Sn particles in the same manner as in Example 1, paste. The results of evaluation in the same manner as in Example 12 are shown.
The average particle size of the particles containing Cu in the connection structure of the solder paste prepared in Comparative Example 3 was 0.8 μm in the same manner as in Example 1. The peak value of the particle size distribution was 0. It was 4 μm.
As is apparent from Table 3, it was confirmed that even in the case of Cu alloy particles having the same composition, the voids contained in the solder connection portion were smaller in Example 12 than in Comparative Examples 3 and 4.

[Example 15]
The solder paste prepared in Example 1 was printed on the Cu electrode of the printed circuit board, and after mounting the 1005R component, a reflow heat treatment was performed at a peak temperature of 250 ° C. by the heat treatment method to prepare a sample.
Next, the component joining strength in the shear direction of the produced sample was measured with a push-pull gauge at a pushing speed of 10 mm / min, and the average value of 30 points was defined as the joining strength. Next, after putting it in an oven at 150 ° C. for 1000 hours and measuring the component bonding strength in the same manner as described above, the bonding strength of the 1005R component was reduced from 8.4 N to 7.1 N, and the reduction rate was 15.5%. there were. The results are shown in Table 4 below.

[Examples 16 and 17]
Table 4 below shows the results of evaluating the solder paste produced in Example 4 by the same method as in Example 15 as Example 16.
Table 4 below shows the results of evaluating the solder paste produced in Example 6 by the same method as in Example 15 as Example 17.

[Comparative Examples 7 to 9]
Table 4 below shows the results of a representative lead-free solder Sn-3.0Ag-0.5Cu evaluated as a comparative example 7 in the same manner as in Example 15.
Further, as Comparative Example 8, the result of evaluating the solder paste produced in Comparative Example 3 by the same method as in Example 15 is shown.
Further, as Comparative Example 9, the result of evaluating the solder paste produced in Comparative Example 6 by the same method as in Example 15 is shown.
As is apparent from Table 4, the reduction in bonding strength (reduction rate) in the 150 ° C. high-temperature standing test was larger in Comparative Examples 7 to 9 than in Examples 15 to 17. This is considered to be due to the influence of the CuSn compound grown at the Cu electrode or Cu particle interface. On the other hand, in Examples 15-17, the Cu-Sn-In alloy phase exists and it is thought that the coarse crystallization was suppressed. Note that it is assumed that the initial bonding state is better in Example 15 than in Comparative Example 8 because there are fewer voids.

[Example 18]
The solder paste prepared in Example 1 was printed on the Cu electrode of the printed circuit board, and after mounting 87 pieces of 0603R components (zero ohm resistors) in a series circuit, the peak temperature was 250 ° C. by the heat treatment method. A sample was prepared by reflow heat treatment.
Next, after measuring the circuit resistance of the prepared sample by the 4-terminal method, it was put into a refrigeration cycle tester with -40 ° C and 125 ° C for 30 minutes each as one cycle, and the circuit resistance after 1000 cycles was measured, The resistance change rate was determined to be 0.2%. The results are shown in Table 5 below.

[Examples 19 and 20]
Table 5 below shows the results of evaluation of the solder paste prepared in Example 4 by the same method as in Example 18 as Example 19.
Table 5 below shows the results of evaluation of the solder paste prepared in Example 6 by the same method as in Example 18 as Example 20.

[Comparative Examples 10-12]
Table 5 below shows the results of a typical lead-free solder Sn-3.0Ag-0.5Cu evaluated by the same method as in Example 18 as Comparative Example 10.
Further, as Comparative Example 11, the result of evaluating the solder paste produced in Comparative Example 3 by the same method as in Example 18 is shown.
Further, as Comparative Example 12, the result of evaluating the solder paste produced in Comparative Example 6 by the same method as in Example 18 is shown.
As is apparent from Table 5, the resistance change rate in the cooling / heating cycle test was larger in Comparative Examples 10-12 than in Examples 18-20. This is considered to be due to the influence of the CuSn compound grown at the Cu electrode or Cu particle interface. On the other hand, in Examples 18-20, the Cu-Sn-In alloy phase exists and it is thought that the coarse crystallization was suppressed. Note that it is assumed that the initial bonding state is better in Example 18 than in Comparative Example 11 because there are fewer voids.

  The metal filler, solder paste, and connection structure according to the present invention can be suitably used for solder connection of electronic devices such as component-embedded substrates and packages that undergo multiple heat treatments in the subsequent process.

Claims (6)

A connection structure having particles containing Cu as a component in a conductive region that is lead-free solder,
An alloy phase containing Cu—Sn—In is present at least at the interface of the particles containing Cu, the average particle size of the particles containing Cu is 2 to 28 μm, and the peak position of the particle size distribution is 5 to 5 μm. A connection structure characterized by being 25 μm.   The conductive region is a metal filler composed of 100 parts by mass of Cu alloy particles and 30 to 550 parts by mass of Sn particles or Sn alloy particles, and the Cu alloy particles are composed of Cu and In, and Ag, Sn, and Bi. The connection structure according to claim 1, wherein the connection structure is formed by fusion bonding of a metal filler containing at least one metal selected from the group consisting of the Cu alloy particles having an average particle diameter of 10 to 30 μm.   The connection structure according to claim 2, wherein the Cu alloy particles include 5 to 15 mass% of Ag, 2 to 8 mass% of Bi, 49 to 81 mass% of Cu, 2 to 8 mass% of In, and 10 to 20 mass% of Sn.   The Cu alloy particles have at least one metastable alloy phase observed as an exothermic peak in differential scanning calorimetry (DSC) at 230 to 300 ° C. and at least one melting point observed as an endothermic peak at 480 to 530 ° C. The connection structure according to claim 2, comprising:   The said Cu alloy particle | grains contain the crystal grain of a Cu-Sn alloy phase or a Cu-In alloy phase, and Ag or Bi exists in the interface of these alloy phases, As described in any one of Claims 2-4. Connection structure. The connection structure according to any one of claims 2 to 5, wherein the conductive region is obtained by reflow heat treatment of a solder paste containing the metal filler and a flux component.