JP7081088B2 - Joint structure of electronic components - Google Patents

Description

The present invention relates to a bonding structure of electronic components.

As a bonding structure for electronic components, Patent Document 1 describes a technique in which a semiconductor laser device is mounted on a submount substrate and the submount substrate is bonded to the bonding surface of a heat sink. As a specific configuration, the joint surface of the heat sink is formed as a rough surface, and is bonded to the rough surface via Ag paste or the like as a conductive material.

In Patent Document 1, it is described that heat radiation and adhesion are improved by regularly roughening the rough surface and forming grooves on the rough surface, and the cross-sectional shape of the rough surface is a triangle, a quadrangle, an arc shape, or the like. Is described.

Further, Patent Document 2 describes a technique of forming an insulating heat-dissipating substrate by forming an insulating film on the surface of a heat conductive substrate and joining a circuit board to the insulating film via a bonding layer. As a specific configuration, a material having good thermal conductivity such as aluminum is used for the heat conductive substrate, a ceramic material is used as the insulating film, and a bonding material such as silver brazing or solder containing titanium is used as the bonding layer.

In particular, Patent Document 2 describes that sufficient bonding strength is obtained by forming a plurality of convex portions on the surface of the insulating film on the bonding layer side.

Japanese Unexamined Patent Publication No. 2005-223083 Japanese Unexamined Patent Publication No. 2011-22251

In the bonding structure of Patent Document 1, since a rough surface is formed on the heat sink, the bonding area is increased and the bonding strength between the heat sink and the submount substrate can be expected to be improved. It was thought that it would cause destruction or peeling.

Further, in the bonding structure of Patent Document 2, since a plurality of convex portions of the insulating layer are formed, the bonding area is increased and the bonding strength is improved as in Patent Document 1, but the bonding at the interface between different materials is performed. Therefore, it was considered that cracks would grow due to the action of thermal stress.

That is, in Patent Documents 1 and 2, since different materials having different physical property values are bonded to each other, when the temperature rises due to energization of an element or a circuit, thermal stress due to a difference in thermal expansion causes peeling to the bonding interface. , May cause cracks.

In particular, the drive temperature of semiconductor devices is increasing due to the adoption of SiC (silicon carbide), GaN (gallium nitride), etc., which can be driven at a higher temperature than Si elements, and as described above, due to the difference in thermal expansion of dissimilar materials. It is easy to imagine that peeling and cracking will be noticeable.

For this reason, there is a demand for a bonding structure in which the bonding adhesion between the base material and the semiconductor is maintained in a high state and a good bonding state is obtained even if the temperature change is large.

A feature of the present invention is a surface shape in which a bonding material is sandwiched between a bonding surface of a base material and a semiconductor to perform bonding, and concave portions and convex portions in which the bonding surfaces are smoothly connected are continuously formed. ,
Of the curve of the region extending from the deepest position of the concave portion to the top of the convex portion, between the contact point of the tangential line having a posture orthogonal to the plane on which the joint surface is formed and the top portion, the deepest position. It has an opening edge at a position equal to the stress intensity factor .
The position of the end wall, which is the outer edge of the joining range of the joining material in the joining state, is set in the region between the opening edge and the deepest position in the direction in which the joining material is present from the deepest position of the recess. There is a point.

In this way, by forming the joint surface of the base material into a "waviness shape" having concave portions and convex portions, the contact area between the base material and the joint material is expanded to improve the joint adhesion, and at the same time, in the peeling direction. On the other hand, since it has an anchor effect, cracks and peeling can be suppressed, and interface destruction can also be suppressed. Further, by making the surface of the base material "waviness", for example, as shown in FIG. 9, the first bonding angle θa and the second bonding angle θb corresponding to each of the first substance A and the second substance B are formed. It is possible to partially reduce the stress intensity factor at the interface by changing the angle with.

Further, the contact portion between the end wall of the joint material and the base material is liable to cause a crack, but since the end wall of the joint material is brought into contact with the valley portion of the "waviness shape", the stress representing the driving force of the crack growth is exhibited. It is possible to suppress an increase in the expansion factor. By setting the position of the end wall of the joint member in the valley in this way, the increase in the stress intensity factor is suppressed, and moreover, the stress intensity factor can be maintained below a predetermined small value by setting the position of the end wall. It is possible, and the rationale is explained below.

As shown in FIG. 7, a concave portion 2Sa is formed on the joint surface 2S of the base material 2, and a portion of the convex portion 2Sb is formed on both sides connected to the concave portion 2Sa. With a plurality of positions of the end wall 4W when the materials 4 are joined, the stress intensity factor of each set position is calculated by CAE analysis, and the calculation result is shown in the graph of FIG.

As can be understood from FIG. 8, the stress intensity factors corresponding to a plurality of set positions of the end wall 4W are different, and the stress intensity factor is reduced at the position where the recess 2Sa intersects at the deepest position Px (position of “14” in FIG. 7). It is in a state. By changing the set position of the end wall 4W in the existing direction of the joining material 4 with reference to this deepest position Px, the stress intensity factor is further lowered, and the position where the end wall 4W coincides with the tangent line of the recess 2Sa (in FIG. 7 “. It becomes the minimum at the position of "21").

Further, it can be understood that when the position of the end wall 4W is set in the existing direction of the joining material 4 (left side in FIG. 7) beyond the tangential position Py, the tendency of increasing the stress intensity factor increases. As shown in FIGS. 7 and 8, the intermediate position Pm between "21" and "28" is used as the opening edge. As a result, the end wall 4W, which is the outer edge of the joining range of the joining material 4, is placed inside the opening edge of the recess 2Sa in the direction in which the joining material 4 exists with respect to the deepest position Px of the recess 2Sa. It is possible to suppress an increase in the magnification factor.

As a result, by setting the shape of the joint surface of the base material and the position of the end wall of the joint material in the recess of the joint surface, the joint adhesion between the base material and the semiconductor is maintained high even if the temperature change is large. A joint structure was obtained in which a good joint state was obtained.

As another configuration, the end wall may be arranged at a position where the posture of the tangent line tangent to the curve is orthogonal to the plane on which the joint surface is formed.

By arranging the end wall at the position of the tangent line in the posture orthogonal to the plane on which the joint surface is formed in this way, for example, the direction orthogonal to the plane, that is, the direction in which the joint material is separated from the base material. Even when stress acts on the substrate, the concentration of the stress is avoided and the bonding state between the base material and the bonding material is maintained in the best condition.

It is an enlarged sectional view of an electronic component. It is sectional drawing of the joint structure which set the end wall of a joint material at the deepest position of a recess. It is sectional drawing of the joint structure which set the end wall of a joint material at the tangential position of a recess. It is sectional drawing of the joint structure which joined at the flat interface, and the sectional view of the joint structure where a crack occurred. It is sectional drawing of the joint structure joined at the interface which becomes "waviness shape", and the sectional view of the joint structure where a crack occurred. It is a chart which shows the stress intensity factor with respect to the crack distance of two kinds of joint structures. It is a figure which shows the plurality of setting positions of the end wall of a joint material with respect to the recess of a base material. It is the figure which graphed the stress intensity factor for each set position of the end wall with respect to a recess. It is sectional drawing explaining the stress concentration with respect to a plurality of forms of a joint structure. It is sectional drawing of the joint structure of another embodiment (a).

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[Electronic components]
As shown in FIG. 1, a base material 2 having a circuit pattern of a conductor and a terminal material 3 having a terminal pattern of a conductor are formed on the surface of the insulating substrate 1, and the bonding surface 2S on the surface of the base material 2 is formed. On the other hand, the semiconductor 5 is supported by the die bonding technique via the bonding material 4, and the semiconductor 5 and the terminal material 3 are bonded by the bonding wire 6 to form an electronic component.

The figure shows a power module such as an IGBT or a power transistor as an example of an electronic component. At the site where die bonding is performed, the semiconductor 5 is interposed at the bonding surface 2S on the surface of the base material 2 via the bonding material 4. A joint structure is formed by joining the two. Further, heat dissipation is realized by providing the insulating substrate 1 in the heat sink 7.

[Joining structure]
In the joining structure of this embodiment, Ag paste is used as the joining material 4, and it is configured to suppress cracks and peeling between the base material 2 and the joining material 4. It should be noted that the Ag paste film as the bonding material 4 is bonded to the base material 2 in a state of being formed on the bottom surface 5a (lower surface in FIG. 2) of the semiconductor 5. After this, the Ag paste can be cured by heat treatment.

As shown in FIG. 2, by forming a large number of grooves having a groove width on the surface of the joint surface 2S in a parallel posture, the concave portion 2Sa (valley) and the convex portion 2Sb (mountain) are formed with respect to the joint surface 2S. A "waviness shape" is formed in which the parts) are alternately and smoothly continuous. By setting the position of the end wall 4W, which is the outer edge of the joining material 4 on the joining surface 2S, to the deepest position Px of the recess 2Sa as shown in the figure, the space between the joining material 4 and the base material 2 is set. Cracks and peeling are suppressed.

In order to suppress cracks and peeling more satisfactorily, the position of the end wall 4W is inside the deepest position Px (the side where the joining material 4 is present), and the end wall 4W is at the boundary between the concave portion 2Sa and the convex portion 2Sb. Ideally, it should be set at the tangent position Py where the tangent line is vertical as shown in FIG. By setting the end wall 4W at the tangential position Py in this way, the stress intensity factor becomes the minimum as described later.

The electronic components of this embodiment are used without considering the posture, but in this embodiment, as shown in FIGS. 2 and 3, the surface of the semiconductor 5 facing the base material 2 is the bottom surface. It is referred to as 5a, and the vertical direction and the horizontal direction are described according to these figures, and the direction in which the end wall 4W of the joining material 4 is formed is referred to as the vertical direction.

Further, in order to create a "waviness shape", the surface of the base material 2 is subjected to ultrashort pulse laser processing (femtosecond laser processing or picosecond laser processing). Further, as the bonding material 4, a solder material, a nanoparticle bonding material, a brazing material, and an adhesive can be used.

In this embodiment, as an example of the “waviness shape”, as shown in FIG. 3, the distance T between adjacent convex portions 2Sb is set to about 100 μm, and the distance H from the top of the convex portion 2Sb to the bottom of the concave portion 2Sa is 50 μm. It is set to a degree. In particular, the joint surface 2S has a semicircular shape in which the concave portion 2Sa opens upward, and the convex portion 2Sb has a semicircular shape in which the convex portion 2Sb is convex upward and has a radius of curvature equal to that of the concave portion 2Sa. By smoothly connecting these, a "swell shape" is created. In this "waviness shape", adjacent recesses 2Sa are also arranged at intervals T.

From the cross-sectional shape of the joint surface 2S, the central position (position of T / 2) of the interval T is the deepest position Px, and the position of 1/4 of the interval T in the direction in which the joint material 4 exists from this deepest position Px. Is the tangent position Py. At this tangential position Py, the lower end portion of the end wall 4W of the joining material 4 comes into contact with the intermediate position in the vertical direction (half the distance H (H / 2)).

The concave portion 2Sa and the convex portion 2Sb are not limited to a semicircular shape, and may have a shape such as a quadratic curve such as a parabola or a part of an ellipse. Further, the sinusoidal shape is not excluded as the joint surface 2S. It should be noted that the position of the tangent position Py may not be specified on the joint surface 2S having a "waviness shape" different from that of the concave portion 2Sa and the convex portion 2Sb of the present embodiment.

When the semiconductor 5 is fixed to the base material 2, a film of the bonding material 4 is formed on the bottom surface 5a of the semiconductor 5, and the recess 2Sa of the base material 2 and the end wall of the bonding material 4 are used by a die bonding device. An operation is performed in which the positional relationship with 4W is magnified by a camera and photographed, and the semiconductor 5 is set on the joint surface 2S in a state where the apparatus confirms each positional relationship from the photographed image. Since the die bonding apparatus can be positioned with high accuracy, the operation of setting the end wall 4W of the joining material 4 at the tangential position Py is easily performed as described above.

[Reason for suppressing cracks and peeling]
FIG. 4 shows a bonding structure in which the semiconductor 5 is bonded to the flat bonding surface 2S of the base material 2 via the bonding material 4 on the left side, and stress acts in the peeling direction (vertical direction in FIG. 4). The state in which the crack has grown by the crack distance D is shown on the right side. Further, FIG. 5 shows a joining structure in which the semiconductor 5 is joined via the joining material 4 to the joining surface 2S having a “waviness shape” in which the concave portion 2Sa and the convex portion 2Sb which are smoothly connected are continuously formed. The right part shows a state in which the stress acts in the peeling direction and the crack grows by the crack distance D.

In these two types of joint structures, the stress intensity factor when a crack occurs at the interface between the joint material 4 and the joint surface 2S due to stress is calculated by CAE analysis, and the calculation result is shown in the chart of FIG. There is. In FIG. 6, the growth direction of the crack distance D is taken on the horizontal axis, and the stress intensity factor when the crack grows is taken on the vertical axis.

In this chart, when the joint surface 2S is flat, the stress intensity factor changes as shown in the first graph F, which tends to gradually increase with the progress when the crack distance D grows (increases). On the other hand, when the joint surface 2S has a “waviness shape”, the stress intensity factor changes greatly as shown in the second graph U when the crack distance D grows. In the figure, the “waviness shape” in which the concave portion 2Sa and the convex portion 2Sb are continuously formed is shown as the pattern C of the broken line, and the second graph U changes corresponding to the pattern C.

As can be understood from the figure, in the second graph U, the stress intensity factor greatly decreases at the position corresponding to the recess 2Sa in the “waviness shape” in the horizontal axis direction. Since the stress intensity factor at this position is smaller than the stress intensity factor shown in the first graph F, it is presumed that cracks can be suppressed by allowing the end wall 4W of the joining material 4 to exist at the position of the recess 2Sa of the joining surface 2S. can.

From such an estimation, as shown in FIG. 7, a concave portion 2Sa is formed on the joint surface 2S of the base material 2, and a portion of the convex portion 2Sb is formed on both sides connected to the concave portion 2Sa. On the other hand, the positions of the end wall 4W when the joining material 4 was joined were set at a plurality of places, and the stress intensity factor at each set position was calculated by CAE analysis.

That is, in this CAE analysis, the stress intensity factor in the state where the position of the end wall 4W is set at each of the positions shown in FIGS. 0 to 28 in the horizontal direction is calculated, and the calculation result is obtained in FIG. It is shown as graph G. In FIG. 8, the position of the end wall 4W is taken on the horizontal axis, and the stress intensity factor is taken on the vertical axis.

The position indicated by “14” in the third graph G of FIG. 8 (the position of “14” in FIG. 7 as well) is the central deepest position Px in the width direction of the recess 2Sa, and the deepest position Px is the position of the bonding material 4. By arranging the end wall 4W at the position indicated by "21" in the existing direction (left side in FIG. 7) (similarly at the position of "21" in FIG. 7), the stress intensity factor is maintained at a low value.

The position of this "21" is the tangent position Py, and when the end wall 4W is arranged in the existing direction of the joining material 4 (left side in FIG. 7) beyond this tangent position Py, the stress intensity factor tends to increase. It can be understood that it will increase. Further, in this embodiment, the intermediate position Pm between "21" and "28" and the intermediate position Pn between "7" and "0" are used as an opening edge.

In particular, the opening edge at the intermediate position Pm between "21" and "28" is a position where the stress intensity factor shown by "14" in the third graph G becomes the same value as shown in FIG. The opening edge, which is the intermediate position Pn between "7" and "0", is located at the same height as the intermediate position Pm between "21" and "28" in the vertical direction as shown in FIG. Become.

For this reason, the end wall 4W, which is the outer edge of the joining range of the joining material 4, is arranged inside the opening edge (intermediate position Pm) of the recess 2Sa in the direction in which the joining material 4 exists from the deepest position Px of the recess 2Sa. By doing so, the increase in the stress intensity factor is suppressed, and cracks and peeling are satisfactorily suppressed.

The deepest position Px at the center of the recess 2Sa in the width direction is not limited to the center position where the interval T is accurately divided into two, and includes a slight error range. Similarly, the position of the opening edge of the recess 2Sa is not limited to the above-mentioned one intermediate position Pm and the other intermediate position Pn.

〔supplementary explanation〕
In order to consider the action of stress when cracks occur at the interface between the base material 2 and the bonding material 4, the first substance A and the second substance B are flat as shown at the left end of FIG. It is assumed that the outer wall is formed at the interface between the first substance A and the second substance B in a posture orthogonal to the interface, and the position where the interface is exposed is the intersection Q. Further, in this assumption, the angle from the interface to the outer wall surface of the first substance A around the intersection Q is set as the first junction angle θa, and the angle from the interface to the outer wall surface of the second substance B centered on the intersection Q is the second. The joint angle is θb.

As described above, when the first joint angle θa and the second joint angle θb are equal to each other, a very large stress concentration occurs with respect to the intersection Q. The stress field of this stress concentration is called a singular stress field, and when elastic analysis is performed, the stress and strain become infinite and the intersection Q becomes the starting point of fracture.

On the other hand, as shown in the center and the right end of FIG. 9, when the first joint angle θa and the second joint angle θb are different due to the setting of the posture of the interface and the setting of the posture of the outer wall, the stress is stressed. As a result of the disappearance of the singular field and the relaxation of stress concentration, the intersection Q does not become the starting point of fracture.

From this supplementary explanation, the interface between the joint surface 2S and the joint material 4 is curved by forming the concave portion 2Sa. As a result, for example, even when the end wall 4W of the joining material 4 is set at the deepest position Px of the recess 2Sa as shown in FIG. 2, the position where the end wall 4W of the joining material 4 is in contact with the recess 2Sa (in this supplementary explanation). The stress concentration can be relaxed at the position corresponding to the intersection Q of.

Further, when the end wall 4W of the joining material 4 is set at any position in the region extending from the deepest position Px (see FIG. 2) to the tangential position Py (see FIG. 3) in the recess 2Sa, the supplementary explanation will be given. Since the first joint angle θa and the second joint angle θb are different, stress concentration is applied to the position where the end wall 4W of the joint material 4 is in contact with the recess 2Sa (the position corresponding to the intersection Q in this supplementary explanation). Can be relaxed.

[Action and effect of the embodiment]
By setting the surface shape of the joint surface 2S to the "waviness shape" and setting the position of the end wall 4W of the joint material 4 in the recess 2Sa (valley) in this way, for example, the base material 2 and the joint material 4 can be combined with each other. The difference in the coefficient of thermal expansion is large, and even when the temperature change changes significantly, cracks and peeling are suppressed to maintain a good bonding state.

That is, by forming the joint surface 2S of the base material 2 into a "waviness shape" having the concave portion 2Sa and the convex portion 2Sb, the contact area between the base material 2 and the bonding material 4 is expanded and the bonding adhesion is improved at the same time. Since it has an anchor effect in the peeling direction, cracks and peeling can be suppressed, and interface destruction can also be suppressed.

Ideally, the position of the end wall 4W of the joining material 4 with respect to the recess 2Sa is set to the tangential position Py, but the end wall 4W is displaced in the direction in which the joining material 4 exists with reference to the deepest position Px. Since a good joining state can be maintained by setting the positional relationship so as to be performed, it is not necessary to set the end wall 4W of the joining material 4 at the optimum position based on the deepest position Px of the recess 2Sa, and the joining is performed. Facilitates alignment for.

[Another Embodiment]
The present invention may be configured as follows in addition to the above-described embodiments (those having the same functions as those of the embodiments are designated by the same numbers and reference numerals as those of the embodiments).

(A) As shown in FIG. 10, the basic shapes of the concave portion 2Sa and the convex portion 2Sb of the joint surface 2S are set in the same manner as those described in the embodiment, and the concave portions 2Sd having a small diameter are formed on these surfaces. Form a large number of grooves. As a result, a part of the bonding material 4 with respect to the bonding surface 2S is made to penetrate into the concave portion 2Sd to improve the adhesive performance, and cracks and peeling are further suppressed.

As a modification of this other embodiment (a), instead of the concave portion 2Sd, a large number of independent dimples are formed on the surface of the concave portion 2Sa and the convex portion 2Sb of the joint surface 2S, or a large number of grooves are formed. The concave portion 2Sd of the above may be combined with a large number of dimples.

(B) Although the die bonding bonding structure is described in the embodiment, for example, it may be applied to a bonding structure for adhering a part of an electronic component to a substrate or a heat sink as a base material.

(C) In the embodiment, it has been described that cracks and peeling are suppressed for one end wall 4W of the joining material 4, but the joining material 4 has, for example, end walls 4W at both ends in the lateral direction in FIGS. 2 and 3. Therefore, the lateral dimensions of the semiconductor 5 may be set so that the end walls 4W at both end positions are both set at positions where cracks and peeling are suppressed.

In this alternative embodiment (c), it is assumed that the lateral dimension of the bottom surface 5a of the semiconductor 5 and the dimension of the bonding material 4 in the same direction are equal to each other, and the lateral dimension of the bottom surface 5a of the semiconductor 5 is assumed. By setting the dimension of 1 to an integral multiple of the interval T shown in FIGS. 2 and 3, for example, only one end wall 4W is set to the deepest position Px, and the position of the other end wall 4W is also set to the deepest position Px. Can be set to.

The present invention can be used for a joining structure of electronic components.

2 Base material 2S Joint surface 2Sa Concave 2Sb Convex 4 Joint material 4W End wall 5 Semiconductor Px Deepest position

Claims (3)

It is a surface shape in which a bonding material is sandwiched between a bonding surface of a base material and a semiconductor to perform bonding, and concave portions and convex portions in which the bonding surfaces are smoothly connected are continuously formed.
Of the curve of the region extending from the deepest position of the concave portion to the top of the convex portion, between the contact point of the tangential line having a posture orthogonal to the plane on which the joint surface is formed and the top portion, the deepest position. It has an opening edge at a position equal to the stress intensity factor .
The position of the end wall, which is the outer edge of the joining range of the joining material in the joining state, is set in the region between the opening edge and the deepest position in the direction in which the joining material is present from the deepest position of the recess. Joining structure of electronic components. The joint structure for electronic components according to claim 1, wherein the end wall is arranged at a position where the posture of the tangent line in contact with the curve is orthogonal to the plane on which the joint surface is formed. It is a surface shape in which a bonding material is sandwiched between a bonding surface of a base material and a semiconductor to perform bonding, and concave portions and convex portions in which the bonding surfaces are smoothly connected are continuously formed.
A position where the position of the end wall, which is the outer edge of the joining range of the joining material in the joining state, is equal to the stress intensity factor at the deepest position in the recess in the direction in which the joining material is present from the deepest position of the recess. Is set in the area between the opening edge and the deepest position, which is
Assuming a curve connecting the concave portion and the convex portion of the joint surface, the end wall is arranged at a position where the posture of the tangent line in contact with the curve is orthogonal to the plane on which the joint surface is formed. Joining structure of electronic parts.

Images