Copper bonding wire, method for manufacturing copper bonding wire, and semiconductor device

Abstract

The invention provides a copper bonding wire which can improve the cutting performance, the bonding strength and the following performance, and inhibit the stripping and deviation of a metal wire from a tool. A copper bonding wire comprising a copper alloy having a copper purity of 99.99 mass% or more, the copper bonding wire having a grain boundary density of 0.01 [mu] m-1 or more and less than 0.6 [mu] m-1 in a cross-section in a direction perpendicular to the bobbin, the grain boundary density of the cross-section in the direction perpendicular to the bobbin being 0.01 [mu] m-1 or more and less than 0.6 [mu] m-1, and the crystal orientation in the bobbin direction in the cross-section being 0.1 [mu] m-1 or more and less than 0.6 [mu] m-1. A crystal orientation lt having an angular difference of 15 DEG or less with respect to the axial direction; 111gt, 111gt; the orientation ratio divided by the crystal orientation lt having an angular difference of 15 DEG or less with respect to the axial direction; 101gt, 101gt; the value obtained from the orientation ratio is 10 to 650 inclusive, the dynamic hardness is 45 to 90 inclusive, and the elastic modulus in the cross section is 20 to 70 GPa inclusive.

Description

Technical Field

[0001] The present invention relates to a copper bonding wire (hereinafter referred to as "copper metal wire"), and in particular to a copper bonding wire suitable for use in connecting a bus bar of a lithium-ion battery module of a power semiconductor or an electric vehicle with an electrode of a lithium-ion battery through which a large current flows, and a method for manufacturing the same, and a semiconductor device using the copper metal wire. Background Art

[0002] Generally, "semiconductor" is a general term for CPU (central processing unit), memory, etc. whose main functions are "computation" and "storage". Semiconductors are used in consumer devices such as personal computers, smart phones, and televisions. On the other hand, power semiconductors are responsible for driving electric motors, charging batteries, and further supplying power to operate microcomputers and LSIs (large-scale integrated circuits). Power semiconductors are mainly used for changing voltage and frequency, power conversion (converting DC to AC or converting AC to DC), etc. Power semiconductors are also called power semiconductor devices (power semiconductor conductor), power devices, power elements, power semiconductor elements, etc.

[0003] The "inverter" installed in energy conservation home appliances such as air conditioners, refrigerators, and washing machines is a common example of using power semiconductors. The inverter controls the speed of the motor by switching the frequency. By freely changing the speed of the motor, the inverter can reduce the useless movement of the motor, which helps to save energy. On the other hand, air conditioners without inverters sometimes have problems such as insufficient temperature stability and high power consumption because they adjust the room temperature by repeatedly running and stopping the motor. These actions of the inverter are achieved by "switching" that uses power semiconductors (power transistors) to finely switch the connection and disconnection of the current.

[0004] In addition to energy-saving home appliances, power semiconductors are also widely used in transportation fields such as electric vehicles and hybrid vehicles. In the transportation field, power semiconductors such as IGBTs (insulated gate bipolar transistors) are used to convert and control power. IGBTs are composed of bonding wires that connect IGBT chips (power chips) to each other and to external electrodes. In order to pass a large current through the bonding wires used for power semiconductors, a relatively thick aluminum metal wire with a wire diameter of more than 40μm and less than 700μm is used.

[0005] As described above, ordinary semiconductors whose main purpose is to operate at low power and power semiconductors whose main purpose is to control high power are considered to be completely different even if they are the same semiconductors. Only a small current flows through the bonding wire used in ordinary semiconductors. Therefore, in the bonding wire used in ordinary semiconductors, thin metal wires made of Au, Ag, Cu and with a wire diameter of less than 30μm are mostly used. As a bonding method, one side of the metal wire is melted to form a molten ball, and heat, ultrasound and pressure are applied to the molten ball to bond it to the semiconductor chip electrode, etc. In contrast, in the bonding wire used in power semiconductors, a large current needs to flow through it, so as described above, a thick wire diameter of 40μm or more, a relatively low resistance and cheap aluminum metal wire is used. In the bonding method of the bonding wire used in power semiconductors, a method is adopted in which a molten ball is not made, ultrasound and pressure are applied to the metal wire at room temperature, and the metal wire is bonded to the electrode, etc.

[0006] In addition, unlike ordinary semiconductors, in power semiconductors, as described above, a large current is repeatedly turned on and off by a switch, so a bonding wire that can withstand the heat load caused by heating and cooling is required. In addition, since it is used in applications that require extremely high safety, such as automobiles and electric vehicles, a high level of durability and reliability is required. As described above, the bonding wires used in ordinary semiconductors and the bonding wires used in power semiconductors have completely different characteristics and properties, even if they are the same bonding wires.

[0007] Furthermore, in recent years, as a measure to respond to global warming and climate change, an international framework (Paris Agreement) for reducing greenhouse gas emissions after 2020 has been adopted at the International Conference on Reducing Greenhouse Gas Emissions (COP21), and the trend towards a decarbonized society is accelerating around the world. Among them, the automobile industry plays an important role. European countries have proposed a policy of completely banning the sale of internal combustion engine vehicles (gasoline vehicles, diesel vehicles) by 2035 and limiting the sale to electric vehicles (including plug-in hybrid vehicles). In addition, the United States has also set a goal of making 50% of the domestic new car market electric vehicles by 2030. This trend is also the case in emerging countries such as China and India, and Japan is also a policy to switch all new car sales to electric vehicles (including hybrid vehicles) by 2035.

[0008] As the trend toward electric vehicles accelerates, there are problems that hinder the popularization of electric vehicles: a short cruising range (the distance that can be traveled from a fully charged state to a fully exhausted state) and a small number of charging facilities (locations).

[0009] In particular, the high performance of power semiconductors is indispensable to extend the driving range. Electric vehicles can drive by using inverters, which are power semiconductors, to convert the direct current of the battery, which is the energy source, into alternating current to drive the motor.

[0010] Until now, power semiconductors have been made of silicon (Si), but in recent years, power semiconductors made of silicon carbide (SiC: silicon carbide) have been released. Existing Si power semiconductors generate energy losses when power is turned on and when the switch is turned on and off. All of these power losses are released from the inverter in the form of heat (heat generation). In order to suppress this heat, a heat dissipation mechanism is required, which results in a reduction in the living space of electric vehicles, an increase in vehicle weight, and a shortened cruising range.

[0011] By applying SiC power semiconductors to electric vehicles, it is possible to reduce energy losses both when power is turned on and when the switch is turned on and off. Therefore, passive components such as transformers that benefit from the increase in switching frequency (high frequency) can be miniaturized, and power semiconductors represented by inverters can be made small and lightweight. In addition, since the heat dissipation mechanism can be simplified, the overall compactness (lightweight) of the electric vehicle can be achieved, and the high-efficiency operation of the inverter can be used to extend the cruising range on a single charge.

[0012] Furthermore, the popularization of battery charging equipment is essential to the popularization of electric vehicles, but at the same time, the demand for fast charging is increasing. As a result, bonding wires that can withstand the large current caused by fast charging are required.

[0013] In this trend of miniaturization and handling large currents, most of the bonding wires in power semiconductors currently use aluminum as the main material. In addition to power semiconductors, aluminum bonding wires are also used as metal wires to connect the busbars of lithium-ion battery modules of electric vehicles and the electrodes of lithium-ion batteries. Since aluminum has high thermal resistance and electrical resistance, it generates a lot of heat when energized, and the heat dissipation effect is low. Therefore, the use of aluminum metal wires runs counter to the trend of miniaturization of power semiconductors. In addition, the following phenomenon has been reported: at the joints of weak aluminum metal wires, the grains coarsen due to the heat generated by the passage of large currents, and the joints are destroyed.

[0014] Therefore, copper wires have attracted attention as a material to replace aluminum wires. Copper wires have lower thermal resistance and electrical resistance than aluminum wires, thus suppressing heat generation, and have a higher recrystallization temperature, thus being able to withstand large currents.

[0015] On the other hand, the use of semiconductors using gallium nitride (GaN), known as wide-gap semiconductors, is being promoted not only in the electric vehicle industry but also in AC (Alternating Current) adapters used to charge smartphones, laptops, etc. By using wide-gap semiconductors, the size can be reduced to about half of the previous size, achieving ultra-miniaturization.

[0016] In addition, power semiconductors are also playing an important role in the IoT (Internet of Things) that has become a hot topic recently. As home appliances equipped with IoT are becoming smaller, thinner, and more dense year by year, the miniaturization, thinning, and high-density of power semiconductors are further developed, and the bonding space between power chips and bonding wires is becoming smaller. Therefore, the bonding of bonding wires is in a situation where limited space has to be effectively utilized.

[0017] As described above, copper wires suppress heat generation compared to aluminum wires and can withstand high output, but are harder and have a larger elastic modulus than aluminum, and thus are a major obstacle to ultra-miniaturization of power semiconductors.

[0018] For example, the bonding wire is usually bonded to the power chip (first bonding) and then bonded to the external electrode (second bonding) in a roughly linear manner in the same direction. However, due to the ultra-miniaturization of power semiconductors, the second bonding must be performed with the tiny gap around the electrode as the target, and sharp bending and twisting have to be applied to make it bend at an acute angle for bonding. Therefore, the bonding wire needs to have the freedom to bend in order to change the direction (angle) of the metal wire during the bonding operation. That is, when power semiconductors are ultra-miniaturized, flexible tracking becomes a very important characteristic for the bonding wire. In addition, the same problem also exists in the so-called reverse wire bonding in which the external electrode side is used as the first bonding and the power chip side is used as the second bonding. In this regard, since the elastic modulus of the copper metal wire is larger than that of the aluminum metal wire, the tracking property of the copper metal wire is very poor. Therefore, the possibility of the copper metal wire deviating from the wedge tool is higher than that of the aluminum metal wire.

[0019] As described above, the inventors have found that when the metal wire connecting the first joint and the second joint is forcibly twisted or sharply bent, the cutting performance of the metal wire may be adversely affected. In other words, sometimes, due to the sharp bend, the metal wire is only slightly deformed, but the cutter cannot enter the metal wire well, and the metal wire cannot be completely cut, resulting in so-called cutting residue. If cutting residue occurs, the metal wire may be peeled off (lift-off) from the joint before it is pulled off, or although it does not reach peeling, it cannot be cut below the height set by the jointer, and the height limit is activated, resulting in the unfavorable situation of the device stopping.

[0020] In summary, in order to reduce greenhouse gas emissions, it is necessary to switch to electric vehicles, and the top priorities are (A) extending the driving range and (B) increasing the production of fast charging equipment, which are the keys to the popularization of electric vehicles. Regarding (A), the lightweight of the car is required, so the miniaturization and high heat dissipation of power semiconductors are necessary, and therefore the demand for copper bonding wires with low electrical resistance and thermal resistance is increased. Regarding (B), it is also necessary to use copper bonding wires that can withstand large currents and high outputs. As described above, in the copper bonding wires used to promote ultra-miniaturized power semiconductors, there is a problem of "low tracking", and there are problems with the "cutting properties" and "bonding strength" of the metal wires due to limited bonding. The biggest task of the present invention is to solve these three metal wire problems at the same time.

[0021] Patent Document 1 describes a method for providing a copper bonding wire having an improved storage life in the atmosphere by reducing the surface grain boundary density to 0.6 (μm / μm). 2 ) and 1.6(μm / μm 2 ) The copper bonding wire used in semiconductor devices below can improve the shelf life, but there are no issues regarding the cutting and tracking properties of the metal wire.

[0022] Patent Document 2 lists the problem of mass production in the mounting process using copper bonding wires, which requires special bonding conditions due to low bonding properties. As for the reasons for delaying practical promotion, it is recorded that the problem can be solved by setting the total ratio of Cu, Cu2O, CuO, and Cu(OH)2 on the surface of the metal wire measured by X-ray photoelectron spectroscopy to 10%, and making the ratio of the total ratio of CuO and Cu(OH)2 equivalent to divalent Cu (Cu[II]) to the ratio of Cu2O equivalent to monovalent Cu (Cu[I]), that is, Cu[II] / Cu[I], in the range of 0.8 to 12, but it does not involve the problems of the cutting property and tracking property of the metal wire. In addition, Patent Document 2 is an invention of a thin wire bonding wire for high-density mounting of semiconductor devices, which is different from the thick wire bonding wire for power semiconductors, and the characteristics and properties required of the metal wire are also different.

[0023] Patent Document 3 describes a method for providing a Cu alloy bonding wire for semiconductor devices that can satisfy the required performance in high-density LSI applications by making the angle difference of the crystal orientation on the surface of the metal wire with respect to a direction perpendicular to a plane including the central axis of the metal wire 15 degrees or less. <100> Crystal orientation, <110> Crystal orientation, <111> The existence ratio of the crystal orientation is 3% or more and less than 27% in terms of average area ratio, which can solve the problem, but does not involve improving the cutting property and bonding strength and suppressing the metal wire from deviating from the tool. In addition, the invention described in Patent Document 3 is the same as Patent Document 2, which is an invention about a thin wire diameter bonding wire for high-density LSI, which is essentially different from the bonding wire for power semiconductor use. Prior art literature Patent Literature

[0024] Patent Document 1: International Publication No. 2021 / 111908 Patent Document 2: International Publication No. 2021 / 167083 Patent Document 3: International Publication No. 2020 / 059856 Summary of the invention Problems to be solved by the invention

[0025] The present invention is completed to solve the above-mentioned problems, and its purpose is to provide a copper bonding wire (hereinafter also referred to as "copper metal wire"), which can improve the cutting property and bonding strength and inhibit the metal wire from deviating from the tool in the lead bonding for connecting the bus bar of a power semiconductor or lithium-ion battery module with the electrode of a lithium-ion battery. In addition, the object of the present invention is to provide a method for manufacturing copper metal wire and a semiconductor device using the copper metal wire, which can improve the cutting property and bonding strength and inhibit the metal wire from deviating from the tool in wire bonding related to, for example, the connection between the bus bar of the power semiconductor and the battery electrode of the above-mentioned battery module. Means of solving problems

[0026] As a result of repeated and intensive studies, the inventors of the present invention have found that the copper wire of the following embodiment, the copper wire produced by the manufacturing method of the embodiment, and the semiconductor device can simultaneously solve the problems of cuttability, bonding strength, and copper wire deviation from the tool in wire bonding.

[0027] The copper bonding wire of the embodiment is made of a copper alloy having a copper purity of 99.99 mass % or more, and has a grain boundary density of 0.01 μm in a cross section perpendicular to the wire axis of the copper bonding wire. -1 Above and less than 0.6μm -1, among the crystal orientations in the linear axis direction in the cross section, the crystal orientations having an angle difference of 15° or less relative to the linear axis direction <111> The orientation ratio divided by the crystal orientation <101> The value obtained by the orientation ratio is 10 or more and 650 or less, the dynamic hardness in the cross section is 45 or more and 90 or less, and the elastic modulus in the cross section is 20 GPa or more and 70 GPa or less.

[0028] Preferably, the copper alloy contains silver (Ag) in an amount of 5 mass ppm or more and 30 mass ppm or less, phosphorus (P), iron (Fe), silicon (Si), arsenic (As) and antimony (Sb) in an amount of 0 mass ppm or more and 3 mass ppm or less respectively and a total of 15 mass ppm or less, relative to the total amount.

[0029] Furthermore, the wire diameter of the copper wire of the embodiment is preferably 40 μm or more and 700 μm or less.

[0030] In addition, a method for manufacturing a copper metal wire according to an embodiment is a method for manufacturing a copper bonding wire, wherein the copper bonding wire is composed of a copper alloy having a copper purity of 99.99 mass % or more, and the manufacturing method comprises: a step of preparing a copper wire material composed of a copper alloy having a copper purity of 99.99 mass % or more; a step of wire drawing the copper wire material; and a step of performing a final heat treatment on the copper wire material after the wire drawing, wherein a grain boundary density in a cross section of the copper bonding wire in a direction perpendicular to the wire axis is 0.01 μm-1 or more and less than 0.6 μm-1, and in a crystal orientation in the wire axis direction in the cross section, a crystal orientation having an angle difference of 15° or less relative to the wire axis direction <111> The orientation ratio divided by the crystal orientation <101> The value obtained by the orientation ratio is 10 or more and 650 or less, the dynamic hardness in the cross section is 45 or more and 90 or less, and the elastic modulus in the cross section is 20 GPa or more and 70 GPa or less.

[0031] A semiconductor device according to an embodiment is a power semiconductor device, comprising: at least one substrate, at least one semiconductor element on the substrate, an electrode on the surface of the semiconductor element, a semiconductor element-on-substrate on the semiconductor element, at least one circuit pattern on the substrate, and a copper metal wire, wherein the copper metal wire connects one or more selected from the group consisting of electrodes on the surface of the semiconductor element, electrodes on the surface of the semiconductor element and external electrodes, electrodes on the surface of the semiconductor element and one circuit pattern in the circuit pattern, electrodes on the surface of the semiconductor element and terminals, two adjacent circuit patterns in the circuit pattern, one circuit pattern in the circuit pattern and terminals, the substrate and the substrate, and the semiconductor element-on-substrate on the semiconductor element and the circuit pattern, wherein the copper metal wire is a copper bonding wire made of a copper alloy having a copper purity of 99.99% by mass or more, and a grain boundary density of 0.01 μm in a cross section of the copper bonding wire in a direction perpendicular to the wire axis is -1 Above and less than 0.6μm -1 , among the crystal orientations in the linear axis direction in the cross section, the crystal orientations having an angle difference of 15° or less relative to the linear axis direction <111> The orientation ratio divided by the crystal orientation <101> The value obtained by the orientation ratio is 10 or more and 650 or less, the dynamic hardness in the cross section is 45 or more and 90 or less, and the elastic modulus in the cross section is 20 GPa or more and 70 GPa or less. In addition, in this specification, the symbol "to" represents a numerical range that is greater than or equal to a value on the left side of the symbol and less than or equal to a value on the right side. Effects of the Invention

[0032] According to the copper bonding wire of the present invention, the cutting property and bonding strength in wire bonding can be improved, and the wire can be prevented from deviating from the tool. According to the method for manufacturing a copper bonding wire of the present invention, a copper wire can be obtained which can improve cutting properties and bonding strength and suppress wire deviation from a tool in wire bonding related to bonding a bus bar and a battery electrode used in, for example, a power semiconductor or a battery module. According to the semiconductor device using the copper metal wire of the present invention, since lift-off does not occur at the metal wire bonding portion, long-term stability of the bonding can be obtained, and since the good tracking property of the copper metal wire does not cause the adverse condition of wedge bonding, further long-term reliability of the bonding (first bonding and second bonding) can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a cross-sectional view schematically showing the structure of a semiconductor device according to an embodiment. Figure 2This is a cross-sectional view schematically showing the structure of a semiconductor device according to another embodiment. Figure 3 This is a cross-sectional view schematically showing the structure of a semiconductor device according to another embodiment. Figure 4 This is a photograph showing traces of the bonding wire being lifted off from the first bonding portion in the experimental sample in which the bonding portion was lifted off. Figure 5 This is a photograph showing traces of the bonding wire being lifted off from the second bonding portion in the experimental sample in which the bonding portion was lifted off. Figure 6 This is a photograph showing the state of the bonder when a cut wire residue is generated after the second bonding. Figure 7 The grain boundary density is 0.18μm -1 A cross-sectional photograph of a copper wire of an embodiment of the present invention. Figure 8 The grain boundary density is 0.68μm -1 Photo of a cross section of a copper metal wire. Fig. 9 Here is a photo of the wire that has not strayed from the tool. Fig.10 Here is a photo of the wire that has strayed from the tool. Fig.11 This is a schematic diagram showing the structure of a bonding head in a front cutting method and a state immediately after wire bonding. Fig.12 It is a schematic diagram showing the structure of a bonding head in a post-cutting method and the state immediately after wire bonding. Fig.13 This is a diagram schematically showing a portion of a lithium-ion battery module. Fig.14 This is a diagram schematically showing a portion of a lithium-ion battery module according to another embodiment. Fig.15 FIG. 1 is a diagram schematically showing the entire lithium-ion battery module. DETAILED DESCRIPTION

[0034] The copper bonding wire of the embodiment of the present invention is described below. The copper wire of the present embodiment is suitable for connecting the busbar of the power semiconductor and the battery module to the battery electrode. The copper wire of the present embodiment is composed of a copper alloy with a copper purity of 99.99% by mass or more, and the grain boundary density in the cross section perpendicular to the wire axis of the copper wire is 0.01 μm. -1 Above and less than 0.6μm -1, among the crystal orientations in the linear axis direction in the cross section, the crystal orientations in which the angle difference with respect to the linear axis direction is less than 15° <111> The orientation ratio divided by the crystal orientation <101> The value obtained by the orientation ratio is 10 or more and 650 or less, the dynamic hardness in the cross section is 45 or more and 90 or less, and the elastic modulus in the cross section is 20 GPa or more and 70 GPa or less. Hereinafter, the process of repeated experiments until the completion of the present invention and the structure and manufacturing method of the copper metal wire of this embodiment will be described in detail.

[0035] As wire cutting errors, the wire may be lifted off from the joint between the wire and the substrate before the wire itself breaks. Lifting is caused by the joint strength being weaker than the breaking strength of the wire. In addition to the cutting property, the main reason for the lift-off is the weak joint strength of the joint between the wire and the substrate. Methods for improving the joint strength will be described later.

[0036] First, the inventors carefully observed the relationship between the cutting property of the copper wire and the metallographic structure of the cross section perpendicular to the wire axis. As a result, it was found that there is a correlation between the density of the grain boundary and the cutting property. It should be noted that the cross section perpendicular to the wire axis of the copper wire is also referred to as the "cross section" hereinafter.

[0037] Specifically, in order to evaluate the continuous bonding of the above-mentioned multiple experimental samples, after continuous bonding with a wire bonding machine, the samples were divided into experimental samples that were bonded to the end and experimental samples that were stopped halfway. If the reasons for the stopping of the experimental samples that were stopped halfway were investigated in detail, there were two reasons: stopping due to the lifting off of the bonding part; and stopping due to the metal wire being pulled to a set height and not being cut although it was about to be pulled apart (stopping due to exceeding the height limit). Figure 4 This is a photograph showing traces of the bonding wire being lifted off from the first bonding portion in the experimental sample in which the bonding portion was lifted off. Figure 5 This is a photograph showing traces of the bonding wire being peeled off from the second bonding part in the experimental sample where the bonding part was lifted off. Figure 5 In the paper, the scratch of a predetermined width entering the longitudinal direction of the paper is the mark of the cutter blade. When the metal wire is cut by the front cut method (full cut method), it may be damaged if the joining object is a fragile material. Figure 6 This is a photograph showing the state of the bonder when the cut wire residue is left after the second bonding. Figure 6 In the case of a collision, it is considered that the adapter has stopped due to exceeding the altitude limit.

[0038] The inventors carefully observed the crystal structure of the cross section of each experimental sample metal wire subjected to the continuous bonding test and found that the density of the grain boundary affects the cutting property of the metal wire, which results in a significant influence on the continuous bonding property. That is, the following tendency is clarified: if the grain boundary density (=grain boundary length (μm) / area of ​​the metal wire cross section (μm) 2 )) is 0.01 μm -1 Above and less than 0.6μm -1 , the wire bonding device can continue to bond to the end without stopping. If the grain boundary density is less than 0.01μm -1 , the metal wire will overstretch before being cut, and stop due to exceeding the height limit. If the grain boundary density is 0.6μm -1 As described above, the metal wire stops due to lift-off from the bonding portion before being cut.

[0039] Generally, there is a tendency that the smaller the grain size, the longer the total length of the grain boundary, and the larger the grain size, the shorter the total length of the grain boundary, but the shapes of individual grains are diverse, and even for grains of approximately the same size, the total length of the grain boundary varies depending on the grain shape. For example, in the case of polygonal grains, even if the grain size is approximately the same, the grain boundary length varies greatly depending on the number of sides of the polygon. Therefore, in terms of investigating the relationship with cutting properties, the inventors consider that it is more appropriate to use grain boundary density as an indicator rather than grain size. Figure 7 The grain boundary density is 0.18μm -1 A cross-sectional photograph of a copper wire of an embodiment of the present invention, Figure 8 The grain boundary density is 0.68μm -1 A cross-sectional photograph of a copper wire.

[0040] In addition, the length of the grain boundary also varies depending on the wire diameter of the metal wire. Basically, the thicker the metal wire, the longer the grain boundary, and the thinner the metal wire, the shorter the grain boundary. Therefore, in the present invention, the total length of the grain boundary (μm) is divided by the area of ​​the cross section perpendicular to the length direction of the metal wire (μm). 2 ) and the value of the grain boundary density obtained. The method for measuring the grain boundary density will be described later.

[0041] The grain boundary density in the cross section of the copper bonding wire of the present embodiment is 0.01 μm -1 Above and less than 0.6μm -1 This can improve the cutting performance of the copper wire. The grain boundary density is more preferably 0.03 μm. -1 Above and 0.5μm -1 Below, more preferably 0.05 μm -1 Above and 0.3μm -1 the following.

[0042] Although the details of the correlation and mechanism between the cutting property and the grain boundary density are unclear, it can be inferred that in the tensile break after a crack is introduced into a metal wire, the propagation of cracks in a metal wire with a low grain boundary density is easily concentrated on a specific grain boundary, which helps to improve the cutting ease of the metal wire.

[0043] Next, the bonding strength is explained. Aluminum is used for bonding wires of power semiconductors, but copper wires are harder and less likely to be flattened than aluminum wires. Therefore, in wedge bonding, there is a tendency for the bonding area of ​​the bonding portion to become smaller, which leads to a decrease in bonding strength. In order to increase the bonding area, if a large load is applied during bonding and the metal wire is forcibly deformed, if the bonding object is a fragile object such as a semiconductor chip, the possibility of damage increases.

[0044] The easier it is to deform a copper wire, the easier it is to crush it. Even a small load can increase the bonding area, thereby enhancing the bonding strength and preventing chip damage. The inventors studied the index of the ease of crushing and concluded that the dynamic hardness considering the elastic properties of the metal wire is the best index.

[0045] The Vickers hardness (Hv) and Knoop hardness (HK) widely used in the hardness measurement of metal materials are obtained by bringing the indenter into contact with the surface of the sample, applying a test force to form a pit, and calculating the surface area based on the diagonal length of the pit after unloading to obtain the hardness. This method cannot evaluate the properties that reflect the elasticity of the material. When examining the joining action process, the metal wire is actually pressed by a wedge-shaped tool, and ultrasonic waves are applied to the wire in a state of elastic deformation, so it is considered not suitable as an indicator to represent the properties of the metal wire.

[0046] Vickers hardness is the most common hardness index. It is a method of evaluating hardness by pressing a material with a diamond indenter and measuring the shape (area) of the indentation formed by the indenter. Vickers hardness is measured by pulling up the indenter after pressing the indenter into the material. Therefore, the material pressed at this time rebounds slightly due to its elastic properties. Therefore, the area of ​​the indentation formed by the indenter is slightly reduced, and the measured hardness value is slightly larger than the actual value.

[0047] Dynamic hardness is a value obtained based on the test force and the depth of the indentation process, and can reflect not only the plasticity of the sample but also the rebound caused by elasticity. Specifically, the inventors believe that the action of a wedge-shaped tool clamping a metal wire and then flattening the metal wire to deform it is similar to a series of actions for dynamic hardness measurement, and is therefore very suitable as a value representing the characteristics of the copper metal wire of the embodiment. In addition, since the dynamic hardness is measured while the indenter is kept pressed into the material, there is no influence of rebound caused by the elasticity of the material. Therefore, the inventors believe that dynamic hardness is the best evaluation method for representing the characteristics of materials with a large elastic modulus, especially copper metal wire.

[0048] In addition, when measuring dynamic hardness, the elastic modulus of the metal wire can also be measured by loading and unloading the test force. The inventors believe that the elastic modulus is also an important parameter indicating the crushing ease and followability of the metal wire, and have added it to the evaluation items.

[0049] The inventors measured the dynamic hardness and elastic modulus of multiple experimental samples produced by different methods and found that if the dynamic hardness in the cross section of the copper metal wire is greater than 45 and less than 90, and the elastic modulus in the cross section of the copper metal wire is greater than 20 GPa and less than 70 GPa, the metal wire will not be flattened by a small load without causing adverse conditions, thereby increasing the bonding area of ​​the bonding portion and eliminating insufficient bonding strength.

[0050] The dynamic hardness (DH) of the copper wire of the present embodiment is more preferably 50 to 80, and further preferably 55 to 70. When the dynamic hardness is less than 45, the bonding area is excessively expanded when the wire is flattened, and the wire contacts the adjacent wire, which increases the possibility of causing a short circuit. When the dynamic hardness exceeds 90, the strength of the wire is too high, and the wire is difficult to flatten. Therefore, not only the bonding area cannot be increased, but also the possibility of causing damage when the bonding object is a semiconductor chip increases.

[0051] The elastic modulus in the cross section of the copper wire of the present embodiment is more preferably 30 GPa or more and 60 GPa or less, and more preferably 35 GPa or more and 50 GPa or less. When the elastic modulus is less than 20 GPa, the possibility of the metal wire joining from the first joint to the second joint to fall over and causing the so-called tilting failure increases. If the elastic modulus exceeds 70 GPa, the possibility of causing the metal wire to deviate from the wedge-shaped tool described later increases. Ideally, the dynamic hardness and elastic modulus are measured in a cross section perpendicular to the length direction of the metal wire in an area closer to the outer peripheral side than the midpoint of the radius.

[0052] The dynamic hardness and elastic modulus can be measured, for example, by attaching a triangular pyramid indenter with an edge angle of 115° to a micro compression tester MCT-W500 manufactured by Shimadzu Corporation under the following conditions. Test mode: loading-unloading test. Test force: 980.665mN. Loading speed: 20.7411mN / sec. Hold time: 10 seconds.

[0053] The determination of the bonding strength is carried out as follows. Using a bonder BJ935 manufactured by HESSE, a tensile test is performed after bonding a metal wire on a copper plate of 50 mm long × 50 mm wide × 1 mm thick. The bonding conditions are set at a load of 2500 gf and a power of 50 V. In each experimental sample, the occurrence rate of peeling of the bonding portion when the central part of the ring is subjected to a tensile test using 20 (n=20) experimental sample metal wires is compared. In the tensile test with n=20, the case where the number of peeling occurs is 3 or more is considered unqualified, and the case where the number of peeling occurs is less than 3 times is considered qualified.

[0054] Next, further problems with bonding wires accompanying the miniaturization of power semiconductors are described. Usually, when sufficient space can be ensured for wire bonding, the bonding action can be performed in a straight line, and the loop curve of the metal wire is not complicated. However, when the area for bonding is limited, since the required number of metal wires are bonded in a limited area, not only the loop curve becomes complicated, but also a bonding action with sharp bending and twisting is performed. For example, in the first bonding (connection with the electrode on the semiconductor chip) and the second bonding (connection with the external electrode on the lead frame or substrate) of the copper metal wire, a wedge-shaped tool (sometimes simply referred to as a "tool") having a mouth (groove portion) for clamping the metal wire at the front end is used, the metal wire is embedded in the mouth, and the metal wire is pressed against the bonding portion, thereby bonding. In a bonding action with many bends and twists, sometimes the metal wire clamped by the wedge-shaped tool (clamp) cannot follow the action and deviates from the mouth of the tool. Copper metal wires with a large elastic modulus (elasticity) are overwhelmingly prone to deviating from the tool compared to aluminum metal wires used in the past.

[0055] If the second bonding is performed while a portion of the metal wire is deviated from the wedge-shaped tool, sufficient bonding cannot be ensured, and depending on the usage environment, the bonded portion may peel off and cause a malfunction. In addition, when the second bonding is performed on a chip, if the bonding operation is performed while the metal wire is deviated from the tool, the tip of the wedge-shaped tool may directly contact the metal film portion on the element, which may cause damage to the semiconductor element.

[0056] By controlling the elastic modulus in the cross section of the copper wire to be within the range of 20 GPa to 70 GPa, the deviation from the tool to a certain extent is eliminated. However, in order to further eliminate the deviation from the tool, the inventors conducted intensive research and found that, in addition to the elastic modulus, by controlling the crystal orientation of the crystals in the cross section perpendicular to the wire axis to have an angle difference of 15° or less with respect to the wire axis direction, the deviation from the tool to the tool can be eliminated. <111> The orientation ratio divided by the crystal orientation <101> The value obtained by controlling the orientation ratio to be greater than 10 and less than 650 can eliminate most of the deviation tool defects. Fig. 9 is a photo of the wire that has not deviated from the tool. Fig. 9 In the photo, the wire is fitted into the crocodile's mouth. Fig.10 is a photo of the wire deviating from the tool. Fig.10 In the photo, the wire deviates from the crocodile's mouth.

[0057] Specifically, the inventors found that: in a cross section perpendicular to the line axis, the crystal orientations in the line axis direction with an angle difference of 15° or less relative to the line axis direction are oriented <111> The orientation ratio is divided by the crystal orientation ratio of the same part under the same conditions <101> The value obtained by setting the orientation ratio (hereinafter also referred to as "crystal orientation ratio ratio") is set to be greater than 10 and less than 650, thereby eliminating most of the deviation from the tool. The above-mentioned crystal orientation ratio ratio in the cross section perpendicular to the wire axis is more preferably greater than 20 and less than 500, and further preferably greater than 30 and less than 450. If the above-mentioned crystal orientation ratio ratio is less than 10, it is not easy to deviate from the tool. If the above-mentioned crystal orientation ratio ratio exceeds 650, it will not deviate from the tool, but the possibility of adverse conditions other than deviation from the tool, such as poor cutting, will increase. The crystal orientation ratio ratio can be calculated using the following formula. Crystal orientation ratio = (crystal orientation <111> Orientation ratio) / (crystal orientation <101> Orientation ratio of

[0058] The mechanism of preventing the tool from deviating by setting the crystal orientation ratio in the cross section perpendicular to the wire axis to 10 or more and 650 or less is not clear, but it is presumed that the crystal orientation <111> The more the crystal orientation is, the more flexible it is to deal with the sudden movement of the tool. In addition, the inventors have successfully found that by finding the crystal orientation of the non-slip surface <101> By the way, although the crystal orientation is also studied, <111> Crystal orientation <100> The crystal orientation ratio was compared with that of the samples, but almost no characteristic tendency was found.

[0059] Furthermore, copper is easily oxidized, and if a copper oxide film is formed on the surface of the copper metal wire, the bonding strength is weakened.

[0060] In addition, as described above, the copper wire of the present invention has a wire diameter of 40 μm to 700 μm, which is thicker than the metal wire (wire diameter 15 μm to 30 μm) used in ordinary semiconductor devices, and the bonding area (volume) is also large, so compared with thin metal wires, a large current can flow at a large voltage. However, if the resistance of the copper wire itself is high, the current is small even at the same voltage, and the advantage of thickness is weakened. The inventors of the present invention believe that: when the electrical conductivity of ultra-high purity copper with less impurities of 99.999% by mass is set to 100%, for the copper wire of this embodiment (copper wire with a purity of 99.99% or more by mass) containing more impurities than it, it is desired to ensure a minimum electrical conductivity of 99% or more as the value when the electrical conductivity of ultra-high purity copper is set to 100%.

[0061] As mentioned above, the popularization of electric vehicles is expected as one of the countermeasures to global warming (reducing carbon dioxide emissions). However, in order to popularize electric vehicles, it is hoped to solve the problem of electric vehicles that the driving distance (cruising range) when fully charged is short compared to gasoline vehicles. In order to extend the driving distance of electric vehicles, miniaturization (lightweighting) of electric vehicles is an important technical issue. In contrast, the more energy loss and the accompanying heat release caused by the resistance of the copper metal wire when the power semiconductor is switched (when turned on and off), the more space is needed for cooling, resulting in the enlargement of electric vehicles. One of the obstacles to the miniaturization of automobiles is the resistance of the copper metal wire. In other words, reducing the resistance of the metal wire is an important key to promoting the popularization of electric vehicles.

[0062] The inventors of the present invention have studied how to adjust the composition of a metal wire made of a copper alloy to simultaneously solve the problem of oxide generation on the surface of the metal wire and the problem of resistance increase. They have found alloy elements that can suppress the generation of oxides and the increase in resistance for copper with a purity of 99.99%. They have conducted in-depth research experiments by adding various elements to the copper matrix and have found that the generation of copper oxides and the increase in resistance can be suppressed by adding silver (Ag), especially by controlling the content of phosphorus (P), iron (Fe), silicon (Si), arsenic (As) and antimony (Sb).

[0063] The present inventors conducted a large number of experiments and found that the content of silver (Ag) is preferably 5 mass ppm or more and 30 mass ppm or less relative to the total amount of the copper alloy. In addition, it was found that phosphorus (P), iron (Fe), silicon (Si), arsenic (As) and antimony (Sb) may not be contained, but the content of each element is preferably adjusted to 3 mass ppm or less, and the total content of phosphorus (P), iron (Fe), silicon (Si), arsenic (As) and antimony (Sb) is 15 mass ppm or less relative to the entire copper metal wire.

[0064] The inventors have discovered that by making the content of silver (Ag) greater than 5 mass ppm and less than 30 mass ppm relative to the total amount of the copper alloy, and simultaneously making the content of each element of phosphorus (P), iron (Fe), silicon (Si), arsenic (As) and antimony (Sb) greater than 0 mass ppm and less than 3 mass ppm, and the total being less than 15 mass ppm, the effect of delaying the formation of copper oxide can be easily obtained, and the overall resistance of the copper metal wire will not increase excessively.

[0065] In particular, in the manufacturing process of the copper metal wire of the present embodiment, if the copper metal wire is exposed to high temperature during heat treatment, the generation rate of copper oxide increases, although it also depends on the state of the heat treatment atmosphere. In this case, the copper metal wire controlled to the above composition has the effect of slowing down the generation rate of copper oxide.

[0066] The progress of oxidation on the surface of the copper wire can be evaluated by measuring the oxide film of copper (II) oxide (CuO) generated on the surface of the copper wire after the copper wire placement test. The placement test is carried out as follows: the copper wire sample is placed in a container without sealing, and is placed at room temperature of 35°C and humidity of 75% RH for up to 10 months. During this period, the CuO oxide film generated on the surface of the copper strip after each period is measured 1 hour after the start of placement, 1 month after 2 months after 3 months after 1 month after 10 months after each period. In the process of copper oxidation, copper (I) oxide (Cu2O) is initially generated, and as the oxidation progresses, the proportion of CuO increases. Therefore, by calculating the proportion of the CuO oxide film, the ease of oxidation generation (oxidation resistance) on the surface of the copper wire can be evaluated. The larger the proportion of the CuO oxide film, the higher the degree of oxidation progress. The proportion of the CuO oxide film can be calculated using the following formula.

[0067] The ratio of CuO oxide film = ((CuO oxide film) / (CuO oxide film + Cu2O oxide film)) × 100 (%)

[0068] In this embodiment, the ratio of the CuO oxide film after 6 months is used as the evaluation standard. If the ratio of the copper (II) oxide (CuO) oxide film is 30% or less, it is evaluated as qualified "A", and if the ratio of the copper (II) oxide (CuO) oxide film is 20% or less, it is evaluated as better "S". The analysis method will be described later.

[0069] In addition, the electrical conductivity adopts the International Annealed Copper Standard (IACS). Compared with the IACS (100%) of 99.999 mass% copper, a sample with a reduction rate of less than 0.5% in IACS means excellent and is designated as "S". A sample with a reduction rate of more than 0.5% and less than 1% in IACS means good and is designated as "A". Details of the IACS measurement method will be described later.

[0070] The wire diameter of the copper wire of the present embodiment is usually 40 μm or more and 700 μm or less, preferably 70 μm or more and 600 μm or less, and more preferably 100 μm or more and 500 μm or less.

[0071] (Method for producing copper metal wire) Next, an example of a method for manufacturing a copper metal wire according to an embodiment is described. It should be noted that the method for manufacturing a copper metal wire is not limited to the method described below. In addition, it is desirable to adjust appropriate conditions according to the weight of the copper metal wire to be manufactured and the processing capacity of the heat treatment furnace.

[0072] In high-purity copper with a purity of 99.99% or more by mass, silver is melted together according to the composition of the copper metal wire to produce a copper melt. The melting can be performed using a heating furnace such as an arc heating furnace, a high-frequency heating furnace, a resistance heating furnace, and a continuous casting furnace. Atmospheric melting is also possible, but in order to prevent the mixing of air in the atmosphere, it is ideal that the copper melt in the heating furnace is melted in a vacuum or an inert gas atmosphere such as argon or nitrogen. The molten material is solidified in the heating furnace in a manner to become a specified wire diameter (diameter) by continuous casting to produce a wire rod. Alternatively, the molten copper can be poured into a mold to produce an ingot, which is placed in an extruder and extruded into a specified wire diameter. In addition, a rough rolled metal wire can also be manufactured by a DIP molding method or an SCR method and used as a wire rod.

[0073] The wire obtained in the above process is drawn into an intermediate wire with a wire diameter of 900 μm. Next, the intermediate wire is subjected to an intermediate heat treatment of heating at 400°C to 600°C for 60 minutes to 420 minutes. The intermediate heat treatment can be a "combustion furnace" using oil, gas, etc. as a heat source, an "electric furnace" using electrical energy as a heat source, and can be "intermittent" or "continuous". In addition, it can also be an "electrical" method in which the metal wire is heated by directly passing electricity. From the viewpoint of preventing oxidation of the metal wire, it is ideal that the heat treatment atmosphere is an inert gas atmosphere.

[0074] Next, the metal wire after the intermediate heat treatment is subjected to wire drawing to the target final wire diameter. In the wire drawing process, the metal wire is passed through a plurality of superhard dies or diamond dies in sequence to reduce the wire diameter of the metal wire in stages. Ideally, the wire reduction rate (also called cross-sectional reduction rate, processing rate) before and after the primary die wire drawing process is 5% or more and 30% or less.

[0075] As a finishing process, the metal wire processed to the final wire diameter is subjected to heat treatment (final heat treatment). The final heat treatment conditions are the most important part for manufacturing the copper metal wire of the embodiment. The inventors have repeatedly tested various heat treatment experiments, and the results show that the following method is the most preferred manufacturing process of the copper metal wire of the embodiment. In the final heat treatment, an electric heating method is adopted. As a condition for electric heating, the voltage value is greater than 16V and less than 28V, the distance between the electrode terminals is greater than 800mm and less than 1300mm, the metal wire moving speed is greater than 50m / minute and less than 200m / minute, and the second electrode is immersed in cooling water (that is, the metal wire moves into the cooling water while being heated). Pure water is used as the coolant, and the temperature of the pure water is greater than 20°C and less than 80°C, and the dissolved oxygen concentration in the pure water is less than 8mg / liter. It should be noted that in order to prevent oxidation of the metal wire, the electrode terminals through which the metal wire passes are covered in a cylindrical shape, and nitrogen is circulated at a flow rate of greater than 30 liters / minute and less than 60 liters / minute. It should be noted that, although the electric heating method is described here, other heat treatment methods may also be used as long as the heating conditions are the same.

[0076] (Semiconductor Device) Next, refer to Figure 1 The configuration of a semiconductor device 100 using a copper wire according to the embodiment will be described.

[0077] like Figure 1 As shown, the semiconductor device 100 includes a semiconductor element 1 , a metal film 2 , a metal wire 3 , a circuit pattern 41 , a metal pattern 42 , an insulating member 43 , a heat dissipating member 5 , a bonding material 6 , a housing 7 , a terminal 8 , and a sealing material 9 .

[0078] In the present embodiment, the semiconductor element 1 is, for example, a power semiconductor used for power supply. Examples of the semiconductor element 1 include a metal oxide semiconductor field effect transistor (MOSFET), an insulated gate bipolar transistor (IGBT), and the like.

[0079] The semiconductor element 1 is formed by sequentially stacking an electrode 11, a substrate 13, and a back electrode 12. The electrode 11 is, for example, an aluminum (Al)-silicon (Si) electrode, and the substrate 13 is, for example, a silicon (Si) substrate, a silicon carbide (SiC) substrate, a gallium nitride (GaN) substrate, or the like.

[0080] The metal film 2 is provided on the surface of the electrode 11 opposite to the substrate portion 13 in a manner covering the surface of the electrode 11. The metal film 2 is a nickel (Ni) film, a copper (Cu) film, a titanium (Ti) film, a tungsten (W) film, etc., and is a film formed by electroplating, chemical plating, evaporation, sputtering, etc. As the nickel (Ni) film, there is a nickel (Ni) chemical plating film, specifically, there can be listed: chemical nickel (Ni)-phosphorus (P) film, chemical nickel (Ni)-boron (B) film, etc. Other preferred embodiments of the metal film 2 will be described later.

[0081] The metal wire 3 is made of the copper metal wire of the above-mentioned embodiment, and its structure and characteristics are also as described above. The metal wire 3 is bonded to the surface of the metal film 2 .

[0082] Next, other structures of the semiconductor device 100 are described. In the semiconductor device 100, a semiconductor circuit is formed by the semiconductor element 1, the metal wire 3, the terminal 8, the circuit pattern 41, and the metal pattern 42. The metal wire 3 is bent in the semiconductor device 100, and the bent portion is bonded to the semiconductor element 1, the terminal 8, the circuit pattern 41, etc.

[0083] In the semiconductor device 100, a bonding material 6, a metal pattern 42, an insulating member 43, a circuit pattern 41, a bonding material 6, and a semiconductor element 1 are sequentially stacked on the surface of the heat dissipation member 5. The bonding material 6 is composed of solder, silver (Ag), etc., and respectively bonds the heat dissipation member 5 to the metal pattern 42, and the circuit pattern 41 to the back electrode 12 of the semiconductor element 1. The insulating member 43 is an insulating substrate, etc.

[0084] The housing 7 is composed of a ring-shaped housing having a space inside, and is arranged so as to surround the outer periphery of the heat dissipation component 5. The semiconductor element 1, the metal film 2, the metal wire 3, the circuit pattern 41, the metal pattern 42, the insulating component 43, the bonding material 6, and the sealing material 9 are accommodated in the internal space of the housing 7.

[0085] The terminal 8 functions as a connection terminal with an external device. The terminal 8 is provided on the upper surface of the housing 7, and is configured in such a manner that one end thereof protrudes from the housing 7 into the internal space of the housing 7, and the other end thereof protrudes from the housing 7 into the external area of ​​the housing 7. The packaging material 9 encloses the semiconductor element 1, the metal film 2, the metal wire 3, the circuit pattern 41, the metal pattern 42, the insulating component 43, and the bonding material 6 and fills the internal space of the housing 7. The sealing material 9 is a gel-like sealing resin, a cured product of a molding resin, or the like.

[0086] Figure 1 The semiconductor device 100 shown may have electrodes 11 on a plurality of semiconductor elements 1, a plurality of circuit patterns 41, and a plurality of terminals 8. The semiconductor device 100 preferably has an electrode-circuit pattern bonding structure, which includes one electrode 11 among the plurality of semiconductor elements 1, one circuit pattern 41 among the plurality of circuit patterns 41, and a metal wire 3 connecting the electrode 11 and the circuit pattern 41. In addition, the semiconductor device 100 preferably has an electrode-terminal connection structure, which includes one terminal 8 among the plurality of terminals 8, one electrode 11 among the plurality of electrodes 11, and a metal wire 3 connecting the terminal 8 and the electrode 11. Furthermore, the semiconductor device 100 preferably has a circuit pattern-terminal connection structure, which includes one terminal 8 among the plurality of terminals 8, one circuit pattern 41 among the plurality of circuit patterns 41, and a metal wire 3 connecting the terminal 8 and the circuit pattern 41. The semiconductor device 100 includes one or more electrode-circuit pattern bonding structures, electrode-terminal connection structures, and circuit pattern-terminal connection structures, respectively, and preferably includes two or more electrode-circuit pattern bonding structures, electrode-terminal connection structures, and circuit pattern-terminal connection structures, respectively. That is, the copper metal wire of the embodiment can be used for connecting the electrode 11 and the circuit pattern 41, the terminal 8 and the electrode 11, or the terminal 8 and the circuit pattern 41.

[0087] The semiconductor device 100 may further include a substrate (substrate on semiconductor element) on the semiconductor element 1. In this case, the semiconductor device 100 preferably includes a bonding structure including: a substrate on semiconductor element, a circuit pattern 41, and a metal wire 3 connecting the substrate on semiconductor element and the circuit pattern 41. The same is true for the semiconductor devices 101 and 103 described later.

[0088] Figure 2 FIG. 1 shows a semiconductor device 101 having a lead frame as another embodiment of a semiconductor device. Figure 2 In the Figure 1 Components of the semiconductor device 100 having the same functions are denoted by the same reference numerals, and detailed description thereof will be omitted. Figure 2 The semiconductor device 101 shown includes a lead frame LF in addition to a semiconductor element 1 , a metal film 2 , a metal wire 3 , an insulating member 43 , a bonding material 6 , and a sealing material 9 . Figure 2 The semiconductor device 101 shown in the figure has a lead frame LF and therefore does not have a housing 7, but may have a housing 7. The lead frame LF is bonded to the surface of the insulating member 43 and has a connection with the housing 7. Figure 1The circuit pattern 41 of the semiconductor device 100 shown in FIG. Figure 2 In the embodiment, the lead frame LF is bonded to the insulating member 43 , but a metal plate (not shown) may be disposed between the lead frame LF and the insulating member 43 .

[0089] The sealing material 9 is provided so as to enclose the semiconductor element 1, the metal film 2, the metal wire 3, the insulating member 43, the bonding material 6, and the lead frame LF. However, the end of the lead frame LF protrudes out of the sealing material 9, and the lead frame LF constitutes the semiconductor element 1 and the metal wire 3 constitutes a circuit, and the protruding end functions as a terminal 8 for connecting to an external device of the semiconductor device 101.

[0090] exist Figure 3 , a semiconductor device 103 having a connecting circuit pattern 41 and a metal wire 3 connecting the circuit pattern 41 is schematically shown as another embodiment of the semiconductor device. The semiconductor device 103 is Figure 1 The semiconductor device 100 shown has another circuit pattern 41 next to the circuit pattern 41, and the metal wire 3 is different from the semiconductor device 100 in that it connects the adjacent circuit patterns 41 to each other, but the other structures are the same. Figure 3 The semiconductor device 103 shown may include a plurality of circuit patterns 41, and may include a bonding structure having a metal wire 3 that connects two adjacent circuit patterns among the plurality of circuit patterns 41. The semiconductor device 103 includes one or more such bonding structures, and preferably includes two or more such bonding structures. That is, the copper metal wire of the embodiment may be used for connecting the circuit patterns 41 to the circuit patterns 41.

[0091] Next, Figure 1 , Figure 2 as well as Figure 3 The manufacturing method of the semiconductor device 100, the semiconductor device 101, and the semiconductor device 103 shown in FIG. 1 is described below. First, the components constituting the semiconductor devices 100 and 101 are prepared, stacked as described above, and joined to each other. Then, the end of the metal wire 3 is joined to the surface of the metal film 2 by ultrasonic bonding or the like. Then, the other end of the metal wire 3 is wedge-bonded to the external electrode ( Figure 1 Terminal 8 or Figure 2 LF). The copper metal wire of the above embodiment is used as the metal wire 3. Then, the sealing resin is injected into the semiconductor device 100 and cured to form the sealing material 9. In the case of the semiconductor device 101, the lead frame carrying the above semiconductor element 1 is arranged in the mold, and the sealing resin is injected and cured to form the sealing material 9.

[0092] In the semiconductor device equipped with the copper wire of the above-mentioned embodiment, since lift-off does not occur at the wire bonding portion, the bonding of the bonding portion 31 can be stably maintained for a long time. Furthermore, due to the good tracking property of the copper wire and the absence of the defect of wedge bonding, a semiconductor device with long-term reliability of bonding (first bonding and second bonding) can be manufactured. Example

[0093] Hereinafter, examples will be described. However, the present invention is not limited to the following examples.

[0094] The copper metal wire of the embodiment is produced according to the following method and conditions. Copper with a purity of 99.99% by mass or more is melted together with silver as an additive element. A continuous casting furnace (heating furnace) is used for melting. In order to prevent the mixing of oxygen from the atmosphere, the copper melt in the heating furnace is evacuated and then melted in a nitrogen atmosphere. The melted material is solidified by continuous casting in the heating furnace in a manner of a predetermined diameter to produce a wire.

[0095] The wire obtained above is drawn to a diameter of 0.9 mm and heated at 500°C for 120 minutes for intermediate heat treatment. After the intermediate heat treatment, it is drawn to a final wire diameter of 0.5 mm (500 μm). The wire drawing process uses multiple superhard dies or diamond dies to reduce the wire diameter in stages. The cross-sectional shrinkage before and after the first wire drawing process is in the range of 5 to 30%.

[0096] The metal wire processed to the final wire diameter is put into an electric heating furnace for final heat treatment. As the conditions for electric heating, the voltage value is above 16V and below 28V, the distance between the electrode terminals is above 800mm and below 1300mm, the metal wire moving speed is above 50m / minute and below 200m / minute, and the second electrode is immersed in cooling water. Thus, the metal wire is immersed in cooling water in a heated state. As a coolant, pure water is used, and electric heating is performed under the conditions that the temperature of the pure water is above 20°C and below 80°C and the dissolved oxygen concentration of the pure water is below 8mg / L. It should be noted that the electrode portion through which the metal wire passes is maintained under a nitrogen atmosphere. The conditions are changed within the above range to obtain samples of Examples 1 to 30. After the final heat treatment, the metal wire is rewound into several spools in units of approximately 200m. The copper purity and some of the elements contained in the copper metal wire of each embodiment are shown in Table 2.

[0097] The copper purity and some of the contained elements of the copper wires of the comparative examples are shown in Table 2. Regarding the manufacturing method of the copper wires of the comparative examples, the manufacturing method from melting to the final wire diameter was carried out under substantially the same conditions as the manufacturing method of the copper wires of the examples, and only the final heat treatment conditions were carried out outside the range of the manufacturing conditions of the examples, and all or part of the above-described range of the electric heating conditions were carried out. Thus, the copper wires of Comparative Examples 1 to 10 were obtained.

[0098] Next, the method of measuring the various properties of the copper wires of Examples and Comparative Examples will be described. (Determination of grain boundary density and crystal orientation ratio) The crystal orientation of the cross section perpendicular to the wire axis direction (the long dimension direction of the wire drawing process) is determined as follows. Prepare a plurality of wire samples that are cut into several centimeters in length from the produced copper wire. While paying full attention to prevent the wire sample from being deformed, stick it straight and flat on the Ag-plated metal frame (metal plate). Then, put the wire sample together with the metal plate into a cylindrical mold (mold) in such a way that the metal plate becomes the bottom surface of the cylinder, pour the embedded resin into the mold, and then add a curing agent to cure the resin. Next, use a grinder to roughly grind the cured cylindrical resin to which the wire sample is added so that the vertical cross section in the long dimension direction of the wire is exposed. Then, the cross-section is finished by final grinding, and then the residual strain of the grinded surface is removed by ion milling to obtain a smooth surface. It should be noted that the ion milling device is finely adjusted in such a way that the cross-section of the wire is perpendicular to the long dimension direction of the wire.

[0099] The grain boundary density and crystal orientation ratio of the copper metal wire of the embodiment and the comparative example were measured and calculated by the backscattered electron analysis (EBSD: Electron Backscattered Diffraction Pattern) method. Specifically, the cross section of the metal wire sample (i.e., the polished surface of the sample) was attached to the sample stage of a field emission scanning electron microscope (FE-SEM: Field Emission Scanning Electron Microscope) equipped with an EBSD measuring device (OIM Data Collection manufactured by EDAX / TSL) in a manner parallel to the sample stage, and the acceleration voltage was set to 15 kV, the measurement point interval was 1.5 μm, the magnification was 170 times, and the boundary with an orientation difference of more than 15° between adjacent measurement points was regarded as a grain boundary, and if the orientation difference was less than 15° and more than 5 pixels were connected, it was recognized as one crystal grain.

[0100] The grain boundary density is calculated by dividing the total length of the grain boundary (μm) by the cross-sectional area of ​​the analysis region (μm 2 ) and calculate the grain boundary density (μm -1It should be noted that the measurement sites were the front end peripheral part, the rear end peripheral part and the middle part of the metal wire of about 200 m, and the grain boundary density in the cross section perpendicular to the long dimension direction of these three sites was measured, and the average value was recorded in Table 1. Similarly, the crystal orientation ratio ratio is also obtained by dividing the crystal orientation in the wire axis direction by the crystal orientation having an angle difference of 15° or less with respect to the wire axis direction in the cross section of the metal wire. <111> The orientation ratio is divided by the crystal orientation ratio of the same part under the same conditions <101> It should be noted that, regarding the measurement location, three locations having the same grain boundary density were measured, and the average value thereof is recorded in Table 1.

[0101] (Determination method of dynamic hardness and elastic modulus) The dynamic hardness and elastic modulus were measured under the following conditions using a triangular pyramid indenter with an edge angle of 115° mounted on a micro compression tester MCT-W500 manufactured by Shimadzu Corporation. Test mode: loading-unloading test. Test force: 980.665mN. Loading speed: 20.7411mN / sec. Hold time: 10 seconds. As for the measurement sample, the front end, rear end and middle part of each metal wire 200m sampled after the final heat treatment were cut at three places, namely the front end, rear end and middle part of the front end and rear end, and buried in the resin in a manner that the cross section in the direction perpendicular to the wire axis is exposed approximately perpendicularly relative to the wire axis. After grinding, the 4 places near the middle point of the radius of the metal wire cross section were measured, and the average value of the measured values ​​at a total of 12 places was recorded in Table 1. The 4 straight lines connecting each of the above 4 parts and the center of the metal wire are respectively in a positional relationship of 90 degrees, and the 4 parts are arranged at approximately equal intervals. It should be noted that, through the above-mentioned measuring device, the dynamic hardness and elastic modulus are measured approximately simultaneously in a series of processes, but the dynamic hardness is determined by the loading of the indenter (depth in the depth direction), and the elastic modulus is a value determined by the rebound of the depressed sample due to the unloading of the indenter. Therefore, they become independent values ​​corresponding to their materials.

[0102] Next, the evaluation method of each sample of Examples and Comparative Examples will be described. (Cutting performance evaluation method) The evaluation was performed by using the BJ935 bonder manufactured by HESSE Company, by measuring the number of errors (stop times) of the device when continuously bonding a metal wire on a copper plate of 50 mm long × 50 mm wide × 1 mm thick. The bonding conditions were set at a load of 2500 gf and a power of 50 V. For each sample, the combination of the first bonding and the second bonding was set to n = 1 (1 group), and continuous bonding was performed in 1000 groups (n = 1000). When the number of errors (stop times) was more than 3 times, it was evaluated as unqualified "X". When the number of errors (stop times) was 2 times, although there were aspects that should be improved, the possibility of practical problems was low, so it was evaluated as "B". When the number of errors (stop times) was 1 time, it meant good and was evaluated as "A". In addition, when the number of errors (stop times) was 0 times, it meant very excellent and was evaluated as "S". The results are shown in Table 1.

[0103] (Joint strength evaluation method) For the bonding strength, a tensile test was also conducted by using a jointer BJ935 manufactured by HESSE, and a metal wire was bonded to a copper plate of 50 mm long, 50 mm wide, and 1 mm thick. The bonding conditions were set to a load of 2500 gf and a power of 50 V. The combination of the first and second bondings was set to n=1 (1 group). The number of times the bonding part peeled off during the tensile test in the center of the loop was compared for 20 groups (n=20) of each sample. In the tensile test of n=20, the case where the number of peeling occurred was 3 or more times and was considered unqualified and evaluated as "X". In the case where the number of peeling occurred was 2 times, although there were aspects that should be improved, the possibility of problems in practical use was low, so it was evaluated as "B". In the case where the number of peeling occurred was 1 time, it meant good and was evaluated as "A". In addition, in the case where the number of peeling occurred was 0 times, it meant very excellent and was evaluated as "S". The results are shown in Table 1.

[0104] (Deviation Tool Evaluation Method) For each sample with a copper wire diameter of 500 μm, an ultrasonic bonding device (wire bonder ASTERION manufactured by K&S) was used to bond the first and second bonds to a copper plate at a distance of 5 mm. The second bonding portion was bent and bonded in the horizontal direction relative to the wire axis direction with a target of a 45° lateral bend. The bonding conditions were set for each sample in such a way that the ultrasonic energy and the applied pressure were the optimal conditions, respectively. It should be noted that the bonding tool used was manufactured by Kulicke&Soffa, model: 127595-20, and the size of the crocodile mouth that clamped the metal wire was an opening (inner diameter) of 0.5 mm, a depth (height) of 0.2 mm, and a length (depth) of 1.0 mm.

[0105] The state of the metal wire in the second bonding part is observed to determine whether the defect of deviation from the tool has occurred. In each sample, the combination of the first bonding and the second bonding is performed 100 times as 1 time. If the metal wire is not bonded or there are more than 4 contact marks with the tool in contact with one side, it is unqualified and set as "X". If there are 2 to 3 contact marks with the tool in contact with one side, although some improvements are desired, there is no problem in practical use, so it is set as "B". If there is 1 contact mark with the tool in contact with one side, it is good, so it is set as "A". If there is no contact mark at all, it is very excellent, so it is set as "S" as an evaluation of deviation from the tool. The evaluation results are shown in Table 1.

[0106] In addition, copper wires of Examples 31 to 111 and Comparative Examples 11 to 30 were obtained in the same manner as in Example 1 and Comparative Example 2 except that the composition and production conditions were changed, and were evaluated in the same manner as in Example 1. The results are shown in Tables 4 to 6.

[0107] (Comprehensive evaluation) When one or more of the three evaluations in the above Tables 1, 4 to 6 is "S" and the others are "S" or "A", it means excellent and is judged as "excellent" in the overall evaluation. When the total of "A" and "S" is 2 or more, it means good and is judged as "good" in the overall evaluation. If there are 2 or more "B" and no "X" evaluation, it means a passing score and is judged as "pass" in the overall evaluation. Even if there is only one sample with "X" evaluation, it means unqualified and is judged as "unqualified" in the overall evaluation, which are recorded in Table 1 respectively. The specific combination of each evaluation (in any order) is as follows. "Excellent": SSS, SSA, SAA. "Good": SAB, SSB, AAB, AAA. “Qualified”: SBB, ABB, BBB. "Not qualified": Even if there is only one X.

[0108] [Table 1]

[0109] Table 2 shows the copper purity and the concentration (content ratio) of some of the contained elements in Examples and Comparative Examples. Each concentration was calculated using an inductively coupled plasma mass spectrometry (ICP-MS) (8800ICP-MS manufactured by Agilent Technologies) owned by Tanaka Electronics Co., Ltd., the applicant.

[0110] Next, the samples of the embodiments and comparative examples were placed at room temperature of 35°C and humidity of 75% RH for 6 months. Then, the oxidation state of the surface of the copper metal wire was analyzed. As a result, the proportion of the oxide film of copper (II) oxide (i.e., CuO) in the sample with an Ag concentration of more than 5 mass ppm was less than 20%. In the process of copper oxidation, copper (I) oxide (i.e., Cu2O) is usually generated initially, and as the oxidation proceeds, the proportion of CuO increases. Therefore, by calculating the proportion of the CuO oxide film, the ease of oxidation generation (oxidation resistance) of the copper metal wire can be evaluated. It should be noted that the proportion of the CuO oxide film is obtained by converting "(CuO oxide film) / (CuO oxide film+Cu2O oxide film)" into a percentage.

[0111] The oxide film is measured by continuous electrochemical reduction analysis (SERA), using the thickness conversion value determined by SERA. The SERA analysis for determining the thickness of the above copper (II) oxide (CuO) and copper (I) oxide (Cu2O) can be performed using QC-200 manufactured by ECI Technology, for example, in the following order. A metal wire with a diameter of 0.5 mm is clamped between two O-rings, and an area with a diameter of 2.1 cm is isolated with a gasket, and a boric acid buffer is injected and saturated with nitrogen. A 150 μA / cm 2 The current density (I) was used to measure the time (seconds) (t) taken for the Cu2O reduction reaction at -0.30V to -0.60V and the CuO reduction reaction at -0.60V to -0.85V. The thickness T (nm) of CuO and Cu2O was calculated based on T = K·I·t using the constant K obtained from Faraday's law. It should be noted that the value of the constant K for CuO is 6.53×10 -5 (cm 3 / A·sec), and the K value of Cu2O is 2.45×10 -4 (cm 3 / A·sec).

[0112] The measurement sites are the front end periphery, the rear end periphery and the middle part of the metal wire, that is, the surface of three sites is measured, and their average values ​​are obtained. Samples with a CuO oxide film ratio of 20% or less are excellent and are set as "S", and samples with a CuO oxide film ratio of more than 20% and less than 30% are good and are set as "A".

[0113] Here, the reason why the ratio of CuO oxide film is judged as good when it is less than 30% is explained. In the power module, IGBT, MOSFET, and diode chips are mounted on a substrate in which a ceramic plate (silicon nitride, aluminum oxide, etc.) and a copper plate are bonded together by active metal bonding (AMB) or direct copper bonding (DCB), and a circuit is formed on the substrate by wire bonding.

[0114] For example, in the case of a copper wire bonded to a copper plate of a DCB substrate, in a temperature cycle test (e.g., -65°C to 150°C) performed in a finished product state sealed with a silicone resin, the thermal expansion and thermal contraction of the silicone resin exerts force on the metal wire bonding portion, which sometimes causes the problem of metal wire peeling. It should be noted that the temperature conditions of the temperature cycle test are determined by the automotive standards used worldwide. The inventors of the present invention believe that the growth of an oxide film on the surface of the copper wire caused by heat is one of the main causes of the peeling. It was also found that if the proportion of the CuO film on the surface of the copper wire after the placement test of the copper wire, that is, after being placed at room temperature of 35°C and humidity of 75% RH for 6 months, is less than 30%, then in the temperature cycle test, the peeling of the bonding portion of the metal wire can be greatly reduced.

[0115] In recent years, the guaranteed operating temperature range of power semiconductors has expanded, for example, there is a requirement to guarantee a range of -65°C to 175°C. In these temperature cycle tests, a higher load is applied to the copper metal wire compared to the above test conditions, which easily causes the metal wire to peel off from the joint. The inventors of the present invention have found that when the ratio of the CuO oxide film on the surface of the metal wire can be made less than 20% in the placement test, such high requirements are met, and even under more severe conditions, the metal wire will not peel off from the joint, and stable bonding strength can be obtained.

[0116] In addition, the inventors found that in addition to the problem of metal wire peeling caused by the formation of copper oxide film, the expansion of the above-mentioned guaranteed temperature range also causes the problem of reduced conductivity at high temperature. That is, the conductivity has a tendency to decrease as the temperature increases. As described above, the copper metal wire of the embodiment has a thicker wire diameter than the wire diameter of the bonding wire of the ordinary semiconductor, so it has the advantage of being able to flow a large current at a high voltage. However, if the resistance increases due to the increase in temperature and the conductivity decreases, the current value may decrease and the driving force may decrease.

[0117] Therefore, in the manufacture of power semiconductor devices, it is necessary to quantitatively predict the increase in resistance of copper bonding wires at higher temperatures. If the resistance increases above this prediction, not only will the designed current value not be achieved, but the heat generated by the metal wires due to the increase in resistance may also have an adverse effect on peripheral components. For example, a serious adverse situation may occur in which the sealing resin as a peripheral component melts due to the heat generated by the metal wires.

[0118] The inventors of the present invention have repeatedly conducted in-depth research experiments to ensure the expansion of the temperature range, especially the increase rate of the resistance of the copper bonding wire in the high temperature range and the heat generation associated therewith. As a result, it was found that if the conductivity (IACS) at room temperature is 99% or more (preferably 99.5% or more), it is within the range of resistance increase generally assumed in the manufacture of power semiconductor devices (within about 1.7 times).

[0119] In order to suppress the increase of resistance and suppress the ratio of CuO oxide film on the surface of copper wire to less than 30%, the inventors repeatedly conducted experiments to evaluate the resistance and the formation of CuO oxide film by adding trace amounts of various elements to high-purity copper. As a result, it was found that the two problems can be solved at the same time by making the copper wire contain 5 mass ppm or more and 30 mass ppm or less of silver (Ag) and phosphorus (P), iron (Fe), silicon (Si), arsenic (As) and antimony (Sb) each 0 mass ppm or more and 3 mass ppm or less, and the total amount of these elements except silver is 15 mass ppm or less.

[0120] The conductivity was measured by IACS (International Annealed Copper Standard) on the same part and the average value was obtained. IACS is a method that uses a four-terminal method with a potential lead distance of 100 mm and a DC current of 200 mA to obtain the resistivity of each metal wire. The resistivity of annealed standard soft copper at room temperature (20°C) is 1.7241×10 -2 μΩm is a value expressed by the ratio when IACS is 100%. Compared with the IACS of 99.999% copper by mass, the sample with a reduction rate of less than 0.5% in IACS means excellent, which is set as "S", and the sample with a reduction rate of more than 0.5% and less than 1% in IACS means good, which is set as "A", as recorded in Table 2. It should be noted that there are cases where the evaluation is "S" and "A" in the comparative examples, but this is a specialized evaluation of the proportion of the CuO oxide film and the reduction rate of IACS, as shown in Table 1, and the other evaluations are "X", so it is judged as "unqualified" in the comprehensive evaluation.

[0121] In addition, the copper wires of Examples 31 to 111 and Comparative Examples 11 to 30 shown in Tables 4 to 6 were also measured for copper purity and the amount of trace elements contained in the same manner as above, and the proportion of the CuO oxide film (CuO ratio) and the IACS reduction rate were evaluated. The results are shown in Tables 7 to 9.

[0122] [Table 2]

[0123] For the same samples as those in the examples and comparative examples described in Table 1, a cutting property evaluation was conducted to determine whether the same cutting property was obtained even when the wire was cut by a back-cut method (also called a back-cut method or a half-cut method). The evaluation was conducted by the same method as in the "cutting property evaluation method" described above, except that the wire was not completely cut (cutting depth was about 50 to 70% of the diameter of the cross section perpendicular to the long dimension of the wire). The results are shown in Table 3.

[0124] [Table 3]

[0125] As can be seen from the results in Table 3, the cutting performance of the wire sample by the post-cutting method was poorer than that of the full-cutting method generally implemented, but the result was at a qualified level. Therefore, it was confirmed that the copper wire of the embodiment can be used without problems even in the post-cutting method. It should be noted that there are cases where the evaluation is "A" in the comparative example, but this is only an evaluation of the cutting performance. As shown in Table 1, it is "X" in the evaluation items other than the cutting performance, and it is also confirmed as "unqualified" in the comprehensive evaluation.

[0126] The front cutting method (also called the full cutting method) and the back cutting method are described. Fig.11 This is a schematic diagram showing the structure of the bonding head 111 in the front-cutting method and the action after the second bonding. The bonding head 111 is a part of the bonder, and is equipped with a bonding tool and a cutter for gripping the metal wire. When the first bonding and the second bonding are performed in sequence, the bonding head 111 moves the gripped metal wire from the first bonding to the second bonding direction, and respectively bonds with the bonding objects. In the bonder of the front-cutting method, a cutter is provided in front of the bonding head. Therefore, after the second bonding, the bonding head (cutter) is moved to the predetermined cutting position of the metal wire. That is, the bonding head is moved along Fig.11 The distance from the second bonding position to the cutting position of the metal wire also depends on the type of bonding head 111, but is about 3 mm. Fig.11As shown, there are obstacles such as the shell 114 of the semiconductor device in the moving direction (direction of the arrow) of the bonding head 111. For example, when the distance from the bonding position to the obstacle is less than 3 mm, the bonding head 111 may collide with the shell 114, causing the bonder to stop, or the semiconductor device to be damaged.

[0127] Fig.12 It is a schematic diagram showing the structure of the bonding head in the back-cutting method and just after the metal wire bonding. The action of the bonding head 111 during the formation of the first bond and the second bond in the back-cutting method is the same as that of the front-cutting method, but the position of the cutter 112 during the second bond is roughly the same as the position of the cutting metal wire 113, so there is no need to move the bonding head (cutter) when cutting the metal wire. Therefore, even if an obstacle such as the housing 114 of the semiconductor device exists near the bonding position, the possibility of the bonding head contacting the obstacle is much smaller than that of the front-cutting method. In the future, with the further miniaturization of semiconductor devices, expectations for the bonding method of the back-cutting method will increase.

[0128] In addition, in the case of reverse bonding (the normal bonding is performed in the order of first bonding on the IC chip and second bonding on the external electrode and circuit pattern on the substrate, but reverse bonding is the opposite, in the order of first bonding, connecting the copper metal wire to the external electrode and circuit pattern, and second bonding, bonding the copper metal wire to the IC chip), the second bonding before cutting the metal wire is on the IC chip. In the front cutting method, i.e., the full cutting method, the blade of the cutter contacts the IC chip, which is particularly vulnerable, and the IC chip may be damaged. Therefore, in reverse bonding, the back cutting method, i.e., not cutting 100% of the metal wire but cutting about 50% of the metal wire and pulling it off, is suitable. Therefore, it can be confirmed that the bonding wire of the present embodiment, which has excellent cutting properties, also exerts the cutting effect not only in the front cutting method but also in the back cutting method.

[0129] [Table 4]

[0130] [Table 5]

[0131] [Table 6]

[0132] [Table 7]

[0133] [Table 8]

[0134] [Table 9]

[0135] Next, the copper wire of the embodiment is applied to the connection between the busbar of the lithium-ion battery module of an electric vehicle and the electrode of the lithium-ion battery. The lithium-ion battery is formed by combining dozens of columnar lithium-ion batteries as a unit, and the gaps between adjacent lithium-ion batteries are filled with a gel-like substance, thereby avoiding contact between adjacent batteries and absorbing vibrations during vehicle operation.

[0136] Fig.13 This is a diagram schematically showing a portion of a lithium-ion battery module. Fig.13 The lithium-ion battery module shown includes a lithium-ion battery 51, a bus bar 52 on the cathode side, a bus bar 53 on the anode side, and a copper wire 50 that electrically connects the lithium-ion battery 51 and the bus bar 52 on the cathode side. Fig.13 The lithium ion battery module shown further includes another copper wire 50 for electrically connecting the lithium ion battery 51 and the anode-side bus bar 53. As the copper wire 50, the copper wire of the above-described embodiment is used. Fig.13 The anode and cathode of the lithium-ion battery module are respectively arranged on the upper surface and the lower surface of the columnar lithium-ion battery, and are usually installed on the vehicle in a manner that the anode faces upward and the cathode faces downward.

[0137] Fig.14 This is a diagram schematically showing a portion of a lithium ion battery module of another embodiment. Fig.14 The lithium-ion battery module shown includes a lithium-ion battery 51, a bus bar 52 on the cathode side, a bus bar 53 on the anode side, and a copper wire 50 that electrically connects the lithium-ion battery 51 and the bus bar 52 on the cathode side. Fig.14 The lithium ion battery module shown further includes another copper wire 50 for electrically connecting the lithium ion battery 51 and the anode-side bus bar 53. As the copper wire 50, the copper wire of the above-described embodiment is used. Fig.14 The lithium-ion battery module shown has both an anode and a cathode disposed at the center and an outer edge of the upper surface of a cylindrical lithium-ion battery, respectively, and is usually mounted on a vehicle with the anode facing upward.

[0138] Fig.15 FIG. 2 is a diagram schematically showing the entire lithium-ion battery module. Fig.15 As shown, the lithium ion battery module houses a plurality of lithium ion batteries 51, and is configured by filling a gel-like substance 54 between the plurality of lithium ion batteries 51. Copper wires 50 are connected to the lithium ion batteries 51.

[0139] In the joining of the busbar and the copper wire in the lithium-ion battery module and the joining of the electrode of the lithium-ion battery and the copper wire (wire bonding), good joining strength can be maintained by using a copper wire with good followability. The reasons are, for example, as follows. Since ultrasonic waves are used in wire bonding, and since the lithium-ion battery is not fixed in the gel-like substance, the lithium-ion battery itself also vibrates due to ultrasonic waves during wire bonding. In addition, there is a height difference between the electrode of the lithium-ion battery and the busbar. Therefore, in wire bonding, a bending operation is performed that takes into account the vibration of the lithium-ion battery and the height difference between the lithium-ion battery and the busbar. If such a copper wire with a high degree of freedom in bending, that is, high followability, is used, it is possible to suppress the undesirable situation that the metal wire deviates from the tool during bonding and the undesirable situation that the joining strength is insufficient, and prevent the copper wire from breaking due to vibration during driving of the car.

[0140] When miniaturization of batteries is required for miniaturization of electric vehicles, copper wires according to embodiments with excellent followability can achieve excellent bonding strength even when complex bending operations are performed in a narrow space. In addition, copper wires according to embodiments with low resistance (high conductivity) are less likely to generate heat, which can improve the safety of lithium-ion batteries that are at risk of explosion under high temperatures (e.g., above 80°C).

[0141] (Evaluation of the bonding properties of battery electrodes and bus bars, etc.) Using the copper wires of Examples 1 to 30 and Comparative Examples 1 to 3, in order to match the actual state of joining the lithium ion battery electrode and the busbar, two copper plates were used, and after the first bonding was performed on the first plate, the second plate set at a position about 20 mm lower than the first plate was subjected to the second bonding. It should be noted that the two plates used for experimental evaluation were not only set with a height difference, but also set in a manner that the positions when the first bonding and the second bonding were projected on a plane were inclined at about 45 degrees, and the lead bonding was performed. Except that the positional relationship between the first bonding and the second bonding was different, the same HESSE company bonder BJ935 as in the above-mentioned embodiment was used under the same conditions as this embodiment to evaluate the cutting property, bonding strength, and tool deviation. As a result, for the samples of Examples 1 to 30, all samples were evaluated as "B" or above. The evaluation of the cutting property of the copper wire of Comparative Example 1 was "X", the evaluation of the bonding strength of the copper wire of Comparative Example 2 was "X", and the evaluation of the tool deviation of the copper wire of Comparative Example 3 was "X".

[0142] As can be seen from the above, in the wire bonding of the copper metal wire in the power semiconductor device, etc., by controlling its grain boundary density, crystal orientation ratio, dynamic hardness, and elastic modulus within the specified range, the following problems can be solved simultaneously: good cutting performance, no metal wire lift-off from the wedge-shaped bonding portion, and metal wire deviation from the wedge tool caused by improved tracking of the metal wire for complex bonding actions such as sharp lateral bending.

[0143] In addition, the copper metal wire of the embodiment has excellent followability, a low generation rate of CuO oxide film, high bonding strength, high electrical conductivity, and is not easy to generate heat. Therefore, it is not only suitable for power semiconductor applications, but also very suitable for joining the bus bar of the lithium-ion battery module of an electric vehicle and the electrode of the lithium-ion battery.

[0144] The copper metal wire of this embodiment can make a great contribution to the development of the automobile industry, power electronics industry, electric railway, electric power industry, etc., and further play a great role in reducing greenhouse gas emissions and preventing global warming.

Claims

1. A copper bonding wire, characterized in that It is a copper bonding wire composed of a copper alloy with a copper purity of 99.99 mass % or more. The grain boundary density of the copper bonding wire in the cross section in the direction perpendicular to the wire axis is 0.01 μm -1 Above and less than 0.6μm -1 , Among the crystal orientations in the linear axis direction in the cross section, the crystal orientations having an angle difference of 15° or less with respect to the linear axis direction <111> The orientation ratio is divided by the crystal orientation with an angle difference of 15° or less relative to the linear axis direction. <101> The value obtained by the orientation ratio is 10 or more and 650 or less, The dynamic hardness in the cross section is 45 or more and 90 or less, And the elastic modulus in the cross section is greater than or equal to 20 GPa and less than or equal to 70 GPa.

2. The copper bonding wire according to claim 1, wherein The copper alloy contains silver Ag in an amount of 5 mass ppm to 30 mass ppm inclusive, phosphorus P, iron Fe, silicon Si, arsenic As, and antimony Sb in an amount of each of 0 mass ppm to 3 mass ppm inclusive, and a total of 15 mass ppm or less.

3. The copper bonding wire according to claim 1 or 2, wherein: The wire diameter is 40 μm or more and 700 μm or less.

4. A method for producing a copper bonding wire, the method comprising: A step of preparing a copper wire material composed of 99.99 mass % or more of a copper alloy; A step of performing wire drawing on the copper wire; and The step of subjecting the copper wire after wire drawing to a final heat treatment, wherein: The grain boundary density of the copper bonding wire in the cross section in the direction perpendicular to the wire axis is 0.01 μm -1 Above and less than 0.6μm -1 , Among the crystal orientations in the linear axis direction in the cross section, the crystal orientations having an angle difference of 15° or less with respect to the linear axis direction <111> The orientation ratio is divided by the crystal orientation with an angle difference of 15° or less relative to the linear axis direction. <101> The value obtained by the orientation ratio is 10 or more and 650 or less, The dynamic hardness in the cross section is 45 or more and 90 or less, And the elastic modulus in the cross section is greater than or equal to 20 GPa and less than or equal to 70 GPa.

5. A semiconductor device, which is a power semiconductor device, comprising: at least one substrate, at least one semiconductor element on the substrate, an electrode on the surface of the semiconductor element, a semiconductor element substrate on the semiconductor element, at least one circuit pattern on the substrate, and a copper metal wire, wherein the copper metal wire connects one or more selected from the group consisting of electrodes on the surface of the semiconductor element, electrodes on the surface of the semiconductor element and external electrodes, electrodes on the surface of the semiconductor element and one of the circuit patterns, electrodes on the surface of the semiconductor element and terminals, two adjacent circuit patterns in the circuit pattern, one of the circuit patterns and a terminal, the substrate and the substrate, and the substrate on the semiconductor element and the circuit pattern, wherein: The copper metal wire is a copper bonding wire composed of a copper alloy having a copper purity of 99.99 mass % or more. The grain boundary density of the copper bonding wire in the cross section in the direction perpendicular to the wire axis is 0.01 μm -1 Above and less than 0.6μm -1 , Among the crystal orientations in the linear axis direction in the cross section, the crystal orientations having an angle difference of 15° or less with respect to the linear axis direction <111> The orientation ratio is divided by the crystal orientation with an angle difference of 15° or less relative to the linear axis direction. <101> The value obtained by the orientation ratio is 10 or more and 650 or less, The dynamic hardness in the cross section is 45 or more and 90 or less, Furthermore, the elastic modulus of the copper bonding wire in the cross section is greater than or equal to 20 GPa and less than or equal to 70 GPa.

Images