Semiconductor device
The semiconductor device addresses damage to power semiconductor elements by using a wiring pattern not in the main current path for initial bonding and a separate control signal pattern, improving reliability and lifespan.
Patent Information
- Application Number
- PCT/JP2026/008541
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-08
- Filing Date
- 2026-03-05
- Publication Date
- 2026-11-12
AI Technical Summary
Existing semiconductor devices suffer from damage to power semiconductor elements during wire bonding, particularly when using high-strength Al wires, due to the load applied during bonding, which affects the mechanical stability and lifespan of the module.
The semiconductor device design includes a wiring pattern that is not in the main current path for one end of the bonding wire, allowing initial bonding to occur on this pattern, reducing the load on the power semiconductor element, and using a separate bonding pattern for control signals that does not transmit current, thereby mitigating damage.
This design reduces damage to power semiconductor elements during wire bonding, enhancing the reliability and lifespan of the module without significant design changes, while maintaining operational functionality.
Smart Images

Figure JP2026008541_12112026_PF_FP_ABST
Abstract
Description
Semiconductor device
[0001] The present invention relates to a semiconductor device having a power semiconductor element.
[0002] In power control apparatuses such as railways and electric vehicles, a semiconductor device having a power semiconductor element (hereinafter sometimes referred to as "power module") plays an important role. The power semiconductor element is a semiconductor element capable of handling a large current, and examples thereof include an insulated gate bipolar transistor (IGBT), a metal oxide semiconductor field effect transistor (MOSFET), and a diode. Since a large current flows through the power semiconductor element, an Al wire having a diameter of, for example, several hundred μm is used as a bonding wire, and the load applied to the element is large. For example, Patent Document 1 describes a power module in which auxiliary bonding wires that support rising portions of a plurality of main bonding wires from below are provided to enhance mechanical stability.
[0003] Japanese Unexamined Patent Application Publication No. 2015-185570
[0004] When an element and a wiring pattern are connected by wire bonding, one end of the wire is bonded on the element, and the other end is bonded on the wiring pattern. At this time, the bonding of the terminal end portion of the wire is often performed on the wiring pattern so that a load at the time of cutting the wire is not applied. When the terminal end portion is disposed on the wiring pattern, the start end portion is disposed on the element, and thus damage to the element during bonding becomes a problem. When the Al wire is changed to a high-strength Al wire to extend the service life of the power module, the damage to the element during bonding is further increased. The same problem also occurs in Patent Document 1. The present invention has been made to solve such a problem, and an object of the present invention is to provide a power module that reduces damage to a power semiconductor element due to wire bonding.
[0005] To solve these problems, the semiconductor device according to the present invention comprises an insulating substrate, a wiring pattern disposed on the substrate, one or more power semiconductor elements disposed on the wiring pattern, and wires connecting the power semiconductor elements to the wiring pattern, wherein one end of the wire connected to the main current electrode of the power semiconductor element is connected to the wiring pattern which is not in the path of the main current, and the other end is connected to the wiring pattern which is in the path of the main current.
[0006] According to the present invention, it is possible to provide a semiconductor device that mitigates damage to power semiconductor elements caused by wire bonding.
[0007] This is a plan view illustrating the schematic of a semiconductor device according to an embodiment. This is a cross-sectional view taken along the line IB-IB in Figure 1A. This is a plan view illustrating the schematic of a semiconductor device according to an embodiment. This is a cross-sectional view taken along the line IIB-IIB in Figure 2A. This is a plan view illustrating the schematic of a semiconductor device according to a modified example. This is a cross-sectional view taken along the line IIIB-IIIIB in Figure 3A. This is a plan view illustrating the schematic of a conventional semiconductor device. This is a cross-sectional view taken along the line IVB-IVB in Figure 4A.
[0008] Embodiments of the present invention will be described with reference to the drawings. In the drawings, some components may be omitted, and the size, shape, and positional relationships of each component may be exaggerated. Directional expressions, including up, down, left, and right, in the description are examples of relative positional relationships and do not limit the direction of use or manufacture. Furthermore, the present invention is not limited to embodiments or modifications, and various design changes are possible based on the claims.
[0009] [Semiconductor Device] The semiconductor device 1 according to this embodiment will be described with reference to Figures 1A to 2B. The semiconductor device 1 is a high-output power module used for power control of automobiles, railway vehicles, etc. As illustrated in Figures 1A and 1B, the semiconductor device 1 comprises a substrate 10, a wiring pattern 20, a power semiconductor element 50, and wires 30. The configurations of the semiconductor device 1 will be described below. Here, the semiconductor device 1 includes a sealing member 60. Figure 1A, a plan view, illustrates the sealing member 60 as seen through the lens, and Figure 1B, a cross-sectional view, illustrates only the outer edge of the sealing member 60. The same applies to the other plan views and cross-sectional views.
[0010] (Substrate) The substrate 10 is an insulating material that forms the basis of the semiconductor device 1. The substrate 10 can be made of ceramics such as silicon nitride. A wiring pattern 20 is arranged on one side of the substrate 10. In this case, a conductive layer 12 of uniform thickness is arranged on the side opposite to the wiring pattern 20. The conductive layer 12 can be in contact with an external heat dissipation member such as a cooling fin to improve the heat dissipation of the semiconductor device 1. The material of the conductive layer 12 can be a metal such as Au, Ag, or Cu, and in this case, it is Cu.
[0011] (Wiring Pattern) The wiring pattern 20 is a component that is placed on the substrate 10 and becomes part of the wiring. The wiring pattern 20 has an element placement pattern 21, a main current pattern 22, a control pattern 24, and a bond pattern 25. The element placement pattern 21 and the main current pattern 22 are wiring patterns that are in the path of the main current of the power semiconductor element 50 and through which the main current flows. The control pattern 24 is a wiring pattern that transmits a control signal to control the power semiconductor element 50.
[0012] The bonding pattern 25 is a wiring pattern that is not in the path of the main current of the power semiconductor element 50 and does not transmit control signals. Therefore, even if the bonding pattern 25 is not connected to the power semiconductor element 50, the semiconductor device 1 can operate normally. Wires 30 having a junction 32 on the bonding pattern 25 have adjacent junctions 32 only on the main current electrode 51A and not on other wiring patterns 20. The junction 32 is the part of the wire 30 that forms a junction with the wiring pattern 20 or the power semiconductor element 50. The bonding pattern 25 can be used as a junction base for the initial bonding of each wire 30 in wire bonding. The material of the wiring pattern 20 can be a metal such as Au, Ag, or Cu, and in this case, it is Cu. Note that the shape and positional relationship of the wiring pattern 20 in the drawing is an example. The element placement pattern 21, main current pattern 22, control pattern 24, and bonding pattern 25 can be provided in accordance with the arrangement of the power semiconductor element 50, etc.
[0013] (Power Semiconductor Element) The power semiconductor element 50 is a semiconductor chip on which transistors, diodes, etc., are formed. Examples of power semiconductor elements 50 include insulated-gate bipolar transistors (IGBTs), metal-oxide-semiconductor field-effect transistors (MOSFETs), diodes, etc. Examples of semiconductor materials that make up the wafer on which the power semiconductor element 50 is formed include silicon (Si) and silicon carbide (SiC). Here, the power semiconductor element 50 consists of an IGBT 51 and a diode 52, with one IGBT 51 and one diode 52 arranged. There are no particular restrictions on the number of IGBTs 51 and diodes 52; for example, there may be three IGBTs 51 and two diodes 52. The IGBTs 51 and diodes 52 are arranged side by side in the left-right direction in the figure and connected by a continuous wire 30. The IGBTs 51 and diodes 52 can be arranged, for example, so that they are electrically in parallel.
[0014] The power semiconductor element 50 is placed on the wiring pattern 20. The power semiconductor element 50 has main current electrodes on its upper and lower surfaces. The side of the power semiconductor element 50 facing the substrate 10 will be referred to as the lower surface, and the side opposite the substrate 10 will be referred to as the upper surface. The electrode on the lower surface of the IGBT 51 is the collector electrode and is joined to the element placement pattern 21 via a bonding member such as solder. The bonding member is omitted in the drawing. The electrode 51A on the upper surface is the emitter electrode and is joined to the wire 30.
[0015] Furthermore, the IGBT 51 has a control electrode 51B on its upper surface to which a control signal is input. The position of the control electrode 51B can be adjusted according to its positional relationship with the wiring pattern 20. The electrode on the lower surface of the diode 52 is the cathode electrode and is joined to the element placement pattern 21 via a bonding member. The electrode 52A on the upper surface is the anode electrode and is connected to the wire 30.
[0016] (Wires) Wires 30 and 40 are components that connect the power semiconductor element 50 to the wiring pattern 20. Here, the wire 30 for the main current and the wire 40 for the control signal are described with different reference numerals. In the drawings, the wire 30 for the main current is shown as a solid line, and the wire 40 for the control signal is shown as a dashed line. The wire 30 for the main current and the wire 40 for the control signal can be the same wire, or they can be different wires. Wires 30 and 40 can be, for example, Al wire, high-strength Al wire made of an alloy of Al and another metal, or Cu wire.
[0017] The main current wire 30 connects electrodes 51A and 52A on the upper surface of the power semiconductor element 50 to the main current pattern 22. The control signal wire 40 connects the control electrode 51B to the control pattern 24. Electrodes 51A and 52A are connected by multiple wires 30. Here, 10 wires 30 are arranged in a row for each IGBT 51 and diode 52. On the other hand, the control electrode 51B is connected by a single wire 40.
[0018] The wires 30 and 40 are joined to the wiring pattern 20 and electrodes at the joint 32. The joining is performed by applying force to the wires 30 and 40 and pressing them against the object to be joined, while applying ultrasonic waves to vibrate them. As a result, the joint 32 is deformed to be wider than the other parts. One wire 30 for the main current is joined to the electrodes 51A and 52A of the power semiconductor element 50 at multiple points. Here, one wire 30 has three joints 32 on the electrode 51A of the IGBT 51 and two joints 32 on the electrode 52A of the diode 52. In addition, one wire 30 has one joint 32 on the main current pattern 22 and one joint 32 on the bond pattern 25, respectively.
[0019] In this manner, in the semiconductor device 1, the wire 30 connected to the main current electrode of the power semiconductor element 50 has one end 30A connected to a wiring pattern (bond pattern 25) that is not in the main current path, and the other end 30B connected to a wiring pattern (main current pattern 22) that is in the main current path. Furthermore, multiple power semiconductor elements 50, in this case two power semiconductor elements 50, an IGBT 51 and a diode 52, are connected between one end 30A and the other end 30B of the wire 30. The control signal wire 40 has one junction 32 on the control electrode 51B of the IGBT 51 and on the control pattern 24, respectively.
[0020] (Sealing Member) The sealing member 60 is a member that protects each component on the substrate 10. The sealing member 60 is positioned so as to be in contact with each component placed on the substrate 10. The material of the sealing member 60 can be, for example, epoxy resin.
[0021] In the semiconductor device 1 having the above configuration, the wire 30 connected to the main current electrode of the power semiconductor element 50 has one end 30A connected to a bonding pattern 25, which is a wiring pattern not in the main current path, and the other end 30B connected to a main current pattern 22, which is a wiring pattern in the main current path. This allows the initial wire bonding to be performed on the wiring pattern 20. This helps to mitigate damage to the power semiconductor element 50 caused by wire bonding. The bonding pattern 25, which is a wiring pattern not in the main current path, only needs to be wide enough to perform the initial wire bonding and has little impact on the characteristics, making it easy to introduce into the wiring pattern. The semiconductor device 1 can improve reliability without making significant design changes.
[0022] As illustrated in Figures 4A and 4B, in a conventional semiconductor device 100, power semiconductor elements 510 and 520 are arranged on a wiring pattern 210. Electrodes 510A and 520A are connected to the wiring pattern 220 by wires 300, and electrode 510B is connected to the wiring pattern 240 by wire 400. In such a conventional semiconductor device 100, one end 300A of wire 300 was located on the power semiconductor element 510. In wire bonding, the ultrasonic vibration load on the object being bonded is large during the first bond. However, after the last bond, an even larger load is applied because the wire is cut. For this reason, the last bond was performed on the wiring pattern. That is, by making one end 300A of wire 300 on electrode 510A the first bond and the other end 300B on the wiring pattern 220 the last bond, damage to the electrode was suppressed.
[0023] However, when the strength of the wire is increased to extend the lifespan of semiconductor devices, damage to the electrodes tends to become more pronounced. For example, if a high-strength Al wire made of an alloy of Al and another metal is used instead of a conventional Al wire, damage to the electrodes at the initial bond becomes more pronounced. The semiconductor device 1 according to this embodiment can avoid the initial bond performed on the electrodes of the power semiconductor element by providing a wiring pattern (bonding pattern 25) that is not in the path of the main current. Furthermore, by performing the bonds at both ends of the wire on the wiring pattern, the reliability of wire bonding can be improved.
[0024] Furthermore, the bond pattern 25 does not necessarily have to be separated for each power semiconductor element 50; it can be a single continuous pattern for power semiconductor elements 50 that are electrically connected in parallel. Also, the transistor is not limited to an IGBT but may be a MOSFET.
[0025] In semiconductor device 1, the power semiconductor element 50 was described as a transistor and a diode. However, if, for example, a diode and a transistor are formed on a single chip, there can be only one power semiconductor element. As illustrated in Figures 2A and 2B, in semiconductor device 1A, one power semiconductor element 50 is connected between one end 30A and the other end 30B of a wire 30. Here, the power semiconductor element 50 is a MOSFET 55 made of SiC, and one wire 30 has three junctions 32 on the electrode 55A. The control electrode 55B is connected to the control pattern 24. There may be one or more MOSFETs 55. If there are multiple, they can be arranged, for example, electrically in parallel.
[0026] Furthermore, in the wire bonding process for semiconductor devices 1 and 1A, the formation of the joint portion 32 of the wire 30 can be carried out from one end 30A to the other end 30B, with one end 30A as the starting end and the other end 30B as the ending end. Alternatively, the formation can be carried out from the other end 30B to the first end 30A, with one end 30A as the ending end and the other end 30B as the starting end. The control electrodes 51B and 55B are covered with a thick oxide film or the like, and the damage from the initial bonding is less compared to the main current electrodes 51A and 55A. For this reason, the wire 40 can be cut after the initial bonding with the control electrodes 51B and 55B at one end 40A, and then bonding with the control pattern 24 at the other end 40B.
[0027] (Modification) Next, a modification of the semiconductor device will be described with reference to Figures 3A and 3B. The semiconductor device 2 according to the modification differs from the semiconductor device 1 according to the embodiment in that one end 40A of the control signal wire 40 is connected to a bond pattern 27, which is a wiring pattern that does not transmit control signals, but otherwise it is the same. So, the control signal wire 40 of the semiconductor device 2 according to the modification will be described. As illustrated in Figures 3A and 3B, the wire 40 connected to the control electrode 53B has one end 40A connected to the bond pattern 27 and the other end 40B connected to a control pattern 24 that transmits control signals. The bond pattern 27 is a wiring pattern that does not transmit control signals. The wire 40 having a bonding portion 32 in the bond pattern 27 has adjacent bonding portions 32 only on the control electrode 53B and not on other wiring patterns 20. For this reason, the semiconductor device 2 can operate normally even when the bond pattern 27 is not connected to the power semiconductor element 50.
[0028] Although the control electrode 53B is covered with a thick oxide film, it still sustains damage from the initial bonding. In the modified semiconductor device 2, the wire 40 connected to the control electrode 53B has one end 40A connected to a bonding pattern 27, which is a wiring pattern that does not transmit control signals, and the other end 40B connected to a control pattern 24, which is a wiring pattern that transmits control signals. This allows the initial wire bonding to be performed on the wiring pattern 20. This further reduces damage to the power semiconductor element 50 caused by wire bonding.
[0029] In the wire bonding process for semiconductor device 2, the formation of the joint portion 32 of the wire 40 can be performed from one end 40A to the other end 40B, with one end 40A as the starting end and the other end 40B as the ending end. Alternatively, the formation can be performed from the other end 40B to the one end 40A, with one end 40A as the ending end and the other end 40B as the starting end. Furthermore, in the wire bonding process, there is an upper limit to the range of bending angles in which the wire can be bent relative to the direction of travel of the tool that formed the previous joint portion 32. Here, for example, the wire can be bent at an angle of 45 degrees relative to the direction of travel immediately before.
[0030] The bonding pattern 27 is separated from the surrounding continuous element placement pattern 21 by a groove 28. The groove 28 is a portion where no members of the wiring pattern 20 are provided, and the inside and outside of the groove 28 are insulated. This allows the semiconductor device 2 to be positioned so that the bonding pattern 27 is embedded in the region of the element placement pattern 21, increasing the degree of freedom in placement. Furthermore, the bonding pattern 27 and the control electrode 53B can be positioned in such a position that the bending angle of the wire 40 is smaller than the upper limit of what is possible. The control electrode 53B can be positioned at the corner of the IGBT 53 close to the bonding pattern 27. Also, as illustrated in Figure 3A, the control electrodes 53B of adjacent IGBTs 53 can be positioned symmetrically with respect to a line passing through the bonding pattern 27. The wires 30 and 40 connected to adjacent IGBTs 53 can also be positioned similarly symmetrically. This reduces the area on which the power semiconductor elements 50 and the wiring pattern 20 are placed.
[0031] In semiconductor devices 1, 1A, and 2, a main current pattern may be added instead of the bond pattern 25. For example, the main current pattern 22, which is the path for the main current, can be divided into two, and one of the divided main current patterns can be placed in the position of the bond pattern 25. This allows the main current pattern to be positioned with the element placement pattern 21 in between. By joining wires 30 to each of the divided main current patterns, both ends of the wires 30 are connected to the wiring patterns in the path of the main current, thereby mitigating damage to the power semiconductor element 50 due to wire bonding. The divided main current patterns can be joined, for example, at the ends of the element placement pattern 21.
[0032] 1 Semiconductor device 2 Semiconductor device (modified) 10 Substrate 12 Conductor layer 20 Wiring pattern 22 Main current pattern (wiring pattern in the main current path) 25 Bond pattern (wiring pattern not in the main current path) 30 Wire (for main current) 30A One end (wire) 30B Other end (wire) 32 Joint 40 Wire (for control signal) 50 Power semiconductor element 51 IGBT 51A Electrode (main current) 51B Control electrode 52 Diode 52A Electrode (main current) 60 Encapsulation material
Claims
1. A semiconductor device comprising: an insulating substrate; a wiring pattern disposed on the substrate; one or more power semiconductor elements disposed on the wiring pattern; and wires connecting the power semiconductor elements to the wiring pattern, wherein, of the wires, the wire connected to the main current electrode of the power semiconductor element has one end connected to the wiring pattern not in the path of the main current, and the other end connected to the wiring pattern in the path of the main current.
2. The semiconductor device according to claim 1, wherein a plurality of power semiconductor elements are connected between one end and the other end of the wire.
3. The semiconductor device according to claim 1, wherein one power semiconductor element is connected between one end and the other end of the wire.
4. The semiconductor device according to claim 1, wherein the power semiconductor element has a control electrode to which a control signal is input, and of the wires, the wire connected to the control electrode has one end connected to the wiring pattern that does not transmit the control signal and the other end connected to the wiring pattern that transmits the control signal.
5. The semiconductor device according to claim 4, wherein the wiring pattern that does not transmit the control signal is separated from the surrounding continuous wiring pattern by a groove.
