Al connecting material

Abstract

Provided is an Al connecting material which is capable of suppressing the occurrence of internal cracks during production and which exhibits excellent temperature cycle reliability. The Al connecting material contains 3.0 mass% to 12.0 mass% of Si, the average value of the ratio (c / d) of the short side length c to the long side length d of the Al phase in the L cross-section (cross-section in the central axis direction including the central axis) of the Al connecting material is 0.25 to 0.7, and the average value of the ratio (e / f) of the short side length e to the long side length f of the Si phase in the L cross-section of the Al connecting material is 0.2 to 0.7.

Description

Technical Field

[0001] The present invention relates to an Al connecting material. Background Art

[0002] In semiconductor devices, bonding wires (wire material) or bonding tapes (strip material) are used to connect electrodes formed on the semiconductor chip to electrodes on the lead frame or substrate. Power semiconductor devices use bonding wires or bonding tapes primarily made of aluminum (Al). The wire diameter of Al bonding wires is typically in the range of 100 μm to 600 μm, while the width of Al bonding tapes is typically in the range of 100 μm to 3000 μm, and the thickness is typically in the range of 50 μm to 600 μm. Here, Al bonding wires and Al bonding tapes are collectively referred to as Al connecting materials.

[0003] The industrial production of Al interconnect materials requires ensuring quality, productivity, and reliability. For example, during the wire drawing process, where Al bonding wire is thinned, defects such as cracks and voids may occur within the wire. Al interconnect materials used in power semiconductor devices are required to exhibit high reliability, as described below, and to suppress the occurrence of defects during their manufacture.

[0004] Power semiconductor devices often use silicon (Si) as the semiconductor chip material, and Al-Si alloys or Al-Cu alloys as the electrodes formed on the semiconductor chips. Furthermore, power semiconductor devices using Al interconnects are often used in high-power equipment such as air conditioners and solar power generation systems, as well as in automotive applications.

[0005] There are two methods for joining Al interconnects: a first method for bonding to electrodes on semiconductor chips and a second method for bonding to electrodes on leadframes or substrates, both of which utilize wedge bonding. Wedge bonding involves applying ultrasonic vibrations and loads to the Al interconnect using a metal fixture (apparatus), breaking the surface oxide films between the Al interconnect and the electrode material, exposing newly formed surfaces and performing solid-phase diffusion bonding. This bonding method is characterized by connecting the interconnect in a solid phase without melting the interconnect, making it a different bonding technique from welding techniques that melt the interconnect.

[0006] Compared to general power semiconductor devices, the next generation of power semiconductor devices is required to operate stably for a long time. Power semiconductor devices repeatedly connect and disconnect current and operate. When current is supplied to the Si semiconductor chip through the Al connecting material, the temperature of the first joint rises. On the other hand, when the current supply is stopped, the temperature of the first joint decreases. In this way, when the power semiconductor is working, the first joint is repeatedly heated and cooled. In this way, the thermal stress caused by the thermal expansion difference between the Al connecting material and the semiconductor chip is repeatedly loaded on the first joint. In the case of using a connecting material composed only of high-purity Al, the thermal stress causes the Al connecting material to be destroyed in a relatively short period of time, making it difficult to meet the performance required by the next generation of power semiconductor devices. Therefore, in the next generation of power semiconductors, it is required to improve the life of the joint accompanied by the heating and cooling of the first joint (hereinafter referred to as "temperature cycle reliability").

[0007] To meet the requirements for thermal cycling reliability, Al bonding materials that focus on improving mechanical strength have been proposed. As a method for improving the mechanical properties of Al bonding materials, a method of adding specific elements to Al has been proposed.

[0008] Patent Document 1 discloses a bonding wire composed of an Al alloy containing at least magnesium (Mg) and silicon (Si), with the combined Mg and Si content being between 0.03% and 0.3% by mass. This patent document discloses that the reduction in the joint strength of the first joint during a cold temperature cycle test in the temperature range of 70°C to 120°C is slowed by the high strength effect of solid solution strengthening of Mg or Si and the cracking inhibition effect of precipitated magnesium silicide (Mg2Si).

[0009] Patent Document 2 discloses a bonding wire characterized by comprising an alloy containing 0.01-0.2% iron (Fe), 1-20 ppm silicon (Si), and the remainder being an aluminum alloy with a purity of 99.997% or greater. The wire exhibits a microstructure with a solid solution content of 0.01-0.06% Fe, a precipitation content of Fe that is no more than seven times the solid solution content of Fe, and an average grain size of 6-12 μm. This patent document discloses that by uniformly dispersing intermetallic compound particles of Fe and Al in the Al, the mechanical strength of the matrix is enhanced, and that the recrystallized particles are further refined, thereby suppressing the reduction in the bonding strength of the first bond during thermal shock testing in the temperature range of -50°C to 200°C.

[0010] Patent Document 3 discloses a bonding wire formed by melting an Al-Si alloy containing 0.1 to 5% silicon (Si) by mass, with the remainder consisting of Al and impurities, then spraying and quenching the melt to form a thin wire. This patent document discloses that quenching the molten Al-Si alloy results in a fine and uniform dispersion of Si, thereby improving mechanical strength.

[0011] Prior art literature

[0012] Patent Literature

[0013] Patent Document 1: Japanese Patent Application Laid-Open No. 2014-131010

[0014] Patent Document 2: Japanese Patent Application Laid-Open No. 2014-129578

[0015] Patent Document 3: Japanese Patent Application Laid-Open No. 59-57440 Summary of the Invention

[0016] Technical problem to be solved by the invention

[0017] As described above, next-generation power semiconductor devices are required to withstand longer periods of use than conventional power semiconductor devices. During operation, the temperature of the first joint repeatedly rises and falls. As a result, the Al interconnect material, which has a larger linear expansion coefficient than the semiconductor chip, generates thermal stress at the first joint due to the difference in linear expansion coefficients between the two, potentially leading to fatigue failure of the Al interconnect material. Temperature cycling tests are one of the tests used to accelerate the evaluation of the lifetime of the joint (temperature cycling reliability) associated with the rising and falling temperatures of the first joint. Al interconnect materials used in next-generation power semiconductors are required to exhibit excellent temperature cycling reliability in temperature cycling tests. However, when using Al interconnect materials that have been strengthened by adding Si, etc., as disclosed in Patent Documents 1 to 3, temperature cycling tests envisioned for use in next-generation power semiconductor devices have revealed a technical problem: cracks develop at a relatively rapid rate within the Al alloy electrode, which has a lower strength than the Al interconnect material, making it difficult to consistently achieve good temperature cycling reliability.

[0018] The existing temperature cycle test (hereinafter also referred to as "TCT") can be easily performed using commercially available experimental equipment. However, since the temperature change rate in TCT is slow, there is a concern about deviation from the faster temperature change rate during the operation of the power semiconductor device. Therefore, recently, in order to approach the conditions of actual use, a high-speed temperature cycle test (hereinafter also referred to as "high-speed TCT") that accelerates the temperature change rate is being studied. The temperature change rate is different from the conventional TCT, which is about 10°C / minute, for example. In the high-speed TCT, the temperature changes at a high speed of about 200°C / minute, for example. Regarding the reliability evaluation of the joint of the Al connecting material, even if the reliability of the Al connecting material is not reduced when evaluated by the existing TCT, if the Al connecting material is evaluated by the high-speed TCT, the joint strength may be reduced and the joint life may be shortened. Therefore, there is a demand for Al connecting materials that show good joint reliability and excellent temperature cycle reliability in the high-speed TCT, which is a more severe test close to the actual use conditions. Hereinafter, the temperature cycle reliability in the high-speed TCT will sometimes be referred to as "high-speed temperature cycle reliability".

[0019] In addition, with regard to Al alloys containing high concentrations of alloying elements, hardening is the cause, resulting in reduced workability, deterioration of manufacturing yield, and reduced quality, which has become an obstacle to practical application. It has also been found that in Al connecting materials that have been strengthened by adding Si, etc., cracks (hereinafter referred to as "internal cracks") are generated in the Al connecting materials during wire drawing. The reason is believed to be the formation of coarse Si crystals, uneven plastic deformation of the Al phase, etc. The existence of internal cracks can induce undesirable situations such as reduced temperature cycle reliability and melting fracture when a large current is applied, so it is required to suppress the occurrence of internal cracks.

[0020] An object of the present invention is to provide an Al connecting material that can suppress the occurrence of internal cracks during production and exhibits excellent temperature cycle reliability.

[0021] Technical means for solving technical problems

[0022] The present inventors conducted in-depth research on the above-mentioned problems and found that the following Al connecting material can solve the above-mentioned problems. Based on this understanding, they further repeated the research and completed the present invention. The Al connecting material is an Al connecting material containing Si in an amount of not less than 3.0 mass% and not more than 12.0 mass%. The average value of the ratio (c / d) of the short side length c to the long side length d of the Al phase in the L cross-section (the cross-section in the direction of the central axis including the central axis) of the Al connecting material is within a specific range, and the average value of the ratio (e / f) of the short side length e to the long side length f of the Si phase in the L cross-section is within a specific range.

[0023] That is, the present invention includes the following contents.

[0024] <1>

[0025] An Al connecting material containing 3.0 mass % or more and 12.0 mass % or less of Si,

[0026] The average value of the ratio (c / d) of the short side length c to the long side length d of the Al phase in the L cross section (cross section in the central axis direction including the central axis) of the Al connecting material is 0.25 or more and 0.7 or less.

[0027] The average value of the ratio (e / f) of the short side length e to the long side length f of the Si phase in the L cross section of the Al interconnector is 0.2 or more and 0.7 or less.

[0028] <2>

[0029] like <1> The Al connecting material,

[0030] The average diameter of the Si phase in the L cross section is 0.8 μm or more and 4 μm or less.

[0031] <3>

[0032] like <1> or <2> The Al connecting material,

[0033] The average diameter of the Al phase in the L cross section is 5 μm or more and 40 μm or less.

[0034] <4>

[0035] like <1> ~ <3> The Al connecting material according to any one of the preceding claims,

[0036] It further contains at least one of Sr, Ca, Na, Ti, and P in an amount of 10 mass ppm to 800 mass ppm in total.

[0037] <5>

[0038] like <1> ~ <4> The Al connecting material according to any one of the preceding claims,

[0039] It further contains at least one of Mn, Sn, Cu, and Zn in a total amount of 10 mass ppm to 500 mass ppm.

[0040] Effects of the Invention

[0041] According to the present invention, it is possible to provide an Al connecting material that can suppress the occurrence of internal cracks during production and exhibits excellent temperature cycle reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1This schematic diagram illustrates the surface (inspection surface) to be measured for Al interconnectors when measuring the shapes (aspect ratio (c / d), aspect ratio (e / f)) and average diameters of Al phases and Si. The surface to be measured is a cross-section (L cross-section) along the central axis of the Al interconnector.

[0043] Figure 2 It is a schematic diagram for explaining the short side length (c) and the long side length (d) of the Al phase in the L-section.

[0044] Figure 3 It is a schematic diagram for explaining the short side length (e) and the long side length (f) of the Si phase in the L cross section.

[0045] Figure 4 This is a schematic diagram for explaining a center chipping defect in the first bonding portion.

[0046] Figure 5 This is an example of internal cracks in an Al bonding material observed using a soft X-ray transmission device. DETAILED DESCRIPTION

[0047] The present invention will be described in detail below based on its preferred embodiments. While reference is sometimes made to the accompanying drawings, each drawing merely schematically illustrates the shapes, sizes, and configurations of the components to facilitate understanding of the invention. The present invention is not limited to the following embodiments and illustrations and may be implemented with modifications as desired without departing from the scope of the present invention and its equivalents.

[0048] [Al connecting material]

[0049] The Al connecting material of the present invention is an Al connecting material containing 3.0 mass % or more and 12.0 mass % or less of Si, and is characterized in that:

[0050] The average value of the ratio (c / d) of the short side length c to the long side length d of the Al phase in the L cross section (cross section in the central axis direction including the central axis) of the Al connecting material is 0.25 or more and 0.7 or less.

[0051] The average value of the ratio (e / f) of the short side length e to the long side length f of the Si phase in the L cross section of the Al interconnector is 0.2 or more and 0.7 or less.

[0052] As mentioned above, in the temperature cycle test (TCT), when using a connection material composed only of high-purity Al, cracks develop at a relatively fast rate inside the connection material, making it difficult to obtain good temperature cycle reliability. On the other hand, when using an Al connection material that has been strengthened by adding Si or the like, it has been confirmed that cracks develop inside the Al alloy electrode with relatively low strength, making it difficult to obtain the temperature cycle reliability required for the next generation of power semiconductor devices. That is, it has been confirmed that in a high-speed temperature cycle test (high-speed TCT) that accelerates the temperature change rate under conditions close to actual use, even Al connection materials whose reliability has not been reduced when evaluated by the existing TCT sometimes experience a reduction in bonding strength and a shortening of bonding life. Furthermore, regarding Al connection materials that have been strengthened by adding Si or the like, internal cracks sometimes occur during their manufacture.

[0053] The present inventors conducted intensive research to address the aforementioned issues and discovered that the following Al connecting material can suppress the generation of internal cracks during its manufacture, exhibits good joint reliability even in a high-speed TCT test, a more severe test close to actual use conditions, and provides excellent temperature cycling reliability. The Al connecting material contains 3.0% to 12.0% Si by mass, has an average value of the ratio (c / d) of the short side length c to the long side length d of the Al phase in the L-section within a specific range, and has an average value of the ratio (e / f) of the short side length e to the long side length f of the Si phase in the L-section within a specific range. The Al connecting material of the present invention significantly contributes to achieving the temperature cycling reliability required of next-generation power semiconductor devices.

[0054] The Al connecting material of the present invention contains Si in an amount ranging from 3.0% to 12.0% by mass, and is composed of an Al phase in which Si is solid-dissolved in the Al, and a Si phase formed by Si crystallization or precipitation. The Al phase may also contain other additive elements in addition to Si. The term "Si phase" collectively refers to Si crystallized materials and Si precipitates. Si crystallized materials are formed from the melt during solidification and are coarse, ranging in size from approximately 1 to 20 μm. In contrast, Si precipitates are formed in a solid state and are relatively small, ranging in size from approximately 0.1 to several μm.

[0055] In the present invention, the L-section of the Al connecting material is a section in the direction of the central axis including the central axis of the Al connecting material, as described in the "(Method for measuring the shape of the Al phase and the Si phase)" column below. Figure 1 As explained.

[0056] The reason why the Al connecting material of the present invention can suppress the occurrence of internal cracks during production and provide excellent temperature cycle reliability is presumed to be as follows.

[0057] First, about internal crack, think that in wire drawing, coarse Si phase becomes the starting point of stress concentration and crack, promotes the growth of crack due to the uneven plastic deformation of Al phase etc. In the processing such as wire drawing, extrusion, Si phase is easily formed into the particle shape of flat on the central axis direction of Al connecting material. Think that by controlling the shape of this Si phase, by relaxing the strain of the periphery of Si phase or making the direction of strain disperse etc., can suppress to produce crack with Si phase as the starting point. In addition, think that by controlling the shape of Al phase, can reduce the tension pressure in the central axis direction of Al connecting material when processing such as wire drawing, rolling, suppress the growth of crack. Like this, infer that utilize the synergistic effect of suppressing the generation of crack and the growth of crack by controlling the shape of Al phase by reducing the starting point of crack by controlling the shape of Si phase, can suppress internal crack when manufacturing.

[0058] Regarding temperature cycling reliability, it is believed that the Si phase has a smaller linear expansion coefficient than Al, which helps to reduce the difference in linear expansion coefficient between the Al interconnect material and the semiconductor chip, thereby reducing the generated thermal stress. Furthermore, the granular Si phase can inhibit the growth of cracks generated at the bonding interface into the Al interconnect material. Furthermore, it is believed that controlling the shape of the Si phase can reduce the thermal strain at the junction between the Al interconnect material and the semiconductor chip.

[0059] As described above, it is speculated that the Al connecting material of the present invention appropriately controls factors contributing to the suppression of internal cracks and the improvement of temperature cycle reliability. As a result, the occurrence of internal cracks during manufacturing can be suppressed and excellent temperature cycle reliability can be achieved.

[0060] -Si concentration-

[0061] By Si concentration being the scope of 3.0 mass % or more and 12.0 mass % or less, it is helpful to reduce the thermal strain of junction, improve temperature cycle characteristics.If less than 3.0 mass %, then improvement effect is small, if exceed 12.0 mass %, then the reduction of initial bonding strength caused by hardening, the damage of semiconductor chip etc. become problem.From the viewpoint of also obtaining good temperature cycle reliability in the high-speed TCT of accelerating temperature change rate close to the conditions of actual use, the concentration of Si in the Al connecting material of the present invention is 3.0 mass % or more, preferably 3.5 mass % or more, more preferably 3.6 mass % or more, 3.8 mass % or more, 4.0 mass % or more, 4.2 mass % or more, 4.4 mass % or more, 4.5 mass % or more, 4.6 mass % or more, 4.8 mass % or more or 5.0 mass %.On the other hand, if the hardness of Al connecting material is too large, then under the bonding condition of generally used ultrasonic vibration or load, easily produce the damage of semiconductor chip when first bonding. From the viewpoint of obtaining good bonding strength when the first bonding is performed under general bonding conditions, the Si concentration in the Al connecting material of the present invention is 12.0 mass% or less, preferably 11.5 mass% or less or 11.0 mass% or less, and more preferably 10.8 mass% or less, 10.6 mass% or less, 10.5 mass% or less, 10.4 mass% or less, 10.2 mass% or less or 10.0 mass% or less.

[0062] The concentration of elements contained in the Al interconnect material of the present invention can be analyzed using, for example, an ICP (Inductively Coupled Plasma) emission spectrometer or an ICP mass spectrometer. If atmospheric pollutants such as oxygen and carbon are adsorbed on the surface of the Al interconnect material, cleaning with an acid or base, depending on the adsorbed substance, prior to analysis is effective.

[0063] - Shape of the Al phase in the L-section -

[0064] It was found that by setting the average value of the ratio of the short side length c to the long side length d of the Al phase in the L-section (c / d) to be in the range of greater than 0.25 and less than 0.7, the generation of internal cracks during manufacturing can be suppressed, and good temperature cycle reliability can also be obtained in high-speed TCT.

[0065] The numerical value of this ratio (c / d) is an indicator of flatness. Figure 2 To explain further. Figure 2 This is a diagram schematically showing the Al phase in the L cross section of the Al connecting material, showing that the central axis direction of the Al connecting material corresponds to Figure 2 The horizontal direction (left and right direction) of the central axis corresponds to Figure 2For the Al phase in the L cross section, the above-mentioned "short side length c" is equivalent to Figure 2 The size indicated by symbol c. In addition, for the Al phase in the L cross section, the above-mentioned "long side length d" is equivalent to Figure 2 The dimension indicated by the symbol d in the figure. Hereinafter, the ratio of the short side length c to the long side length d of the Al phase in the L-section (c / d) will also be referred to as the "Al phase shape ratio (c / d)." The value of the Al phase shape ratio (c / d) can be determined using the grain shape aspect ratio (Grain Shape Aspect Ratio) in the analysis software included with the instrument.

[0066] Furthermore, in order to suppress the generation of internal cracks during manufacturing and to obtain good temperature cycling reliability even in high-speed TCT, the average value of the aspect ratio (c / d) of the Al phase in the L-section only needs to be within the above-mentioned appropriate range, and it is not necessary for the aspect ratio (c / d) of all Al phases to be within the range of 0.25 to 0.7. For example, an Al phase with an aspect ratio (c / d) of less than 0.25 may be included, and an Al phase with an aspect ratio (c / d) exceeding 0.7 may also be included.

[0067] From the viewpoint of being able to further suppress the generation of internal cracks during manufacturing and obtaining further good temperature cycle reliability in high-speed TCT, the average value of the shape ratio (c / d) of the Al phase in the L cross-section of the Al connecting material of the present invention is more preferably greater than 0.3, further preferably greater than 0.32, greater than 0.34 or greater than 0.35, and the upper limit is more preferably less than 0.65, further preferably less than 0.64, less than 0.62 or less than 0.6.

[0068] In the determination of the average value of the shape ratio (c / d) of the Al phase in the L section of the Al connecting material, a method for combining the information of the Al concentration and Si concentration obtained by SEM-EDS with the information of the crystal orientation obtained by backscattered electron diffraction (EBSD) can be used. In detail, in the determination area of the L section of the Al connecting material as the inspection surface, the concentration determination of Al and Si using EDS and the crystal orientation analysis using EBSD are carried out simultaneously. Then, the analysis software attached to the device is utilized to separate and extract the Al phase and Si phase from the measurement result of EDS. Specifically, it is preferred to utilize the function of the analysis software OIM Data Collection (OIM data acquisition) or OIM Anaysis (OIM analysis) (both are TSLSolutions systems) attached to the FE-SEM device, i.e., Chi Scan (chemical auxiliary scanning) function. Then, for the region determined to be the Al phase, the crystal orientation can be analyzed by utilizing the analysis software attached to the device. If the orientation difference between the measurement points is more than 15 °, it is judged to be a grain boundary and the ratio (c / d) is calculated. The average value of the shape ratios (c / d) of each Al phase is defined as the average value of the shape ratios (c / d) of the Al phase. In the process of calculating the shape ratios (c / d) of the Al phase, locations where the crystal orientation cannot be measured, or locations where the orientation analysis is reliable but can be measured, are excluded and calculated. Therefore, in one embodiment, the average value of the shape ratios (c / d) of the Al phase in the L cross section of the Al interconnector of the present invention is calculated by the following steps (1) to (3).

[0069] (1) Using the L-section of the Al interconnector as the inspection surface, the concentrations of Al and Si were measured using EDS and the crystal orientation was measured using EBSD simultaneously.

[0070] (2) Separate and extract Al and Si using the Chi Scan function. Specifically, based on the EDS measurement results for Si, a tolerance corresponding to the Si threshold is set, allowing separation and identification of Al and Si. Crystallographic information for Al and Si from the material file can be used to analyze crystal orientation.

[0071] (3) For the area determined to be the Al phase, the crystal orientation is analyzed. If the orientation difference between the measurement points is 15° or more, it is judged to be a grain boundary and the ratio (c / d) of each grain is calculated. The ratio (c / d) of each grain is averaged to calculate the average value of the shape ratio (c / d) of the Al phase. Here, the average value of the shape ratio (c / d) of the Al phase is calculated using the value of the grain shape aspect ratio (hereinafter referred to as "grain shape aspect ratio") of the analysis software. This value is the average value obtained by averaging the grain shape aspect ratios of each grain. In addition, for the calculation method of the grain shape aspect ratio, the ratio (c / d) of the short side length (c) (Grain Shape Minor Axis: grain shape short axis) and the long side length (d) (Grain Shape Major Axis: grain shape long axis) of one grain is calculated. In addition, for the average calculation, the average value obtained by averaging the area (Area) that can be selected by the software attached to the device is used. By using the average value calculated from the area average, it is possible to accurately measure and determine whether the conditions related to the average value of the Al phase shape ratio (c / d) are met, which is suitable for suppressing internal cracking during manufacturing and achieving good temperature cycling reliability even in high-speed TCT. The area average is calculated by averaging the value obtained by multiplying the ratio of each grain area to the total grain area by the individual grain area value, and the calculation is automated by the software.

[0072] In step (2) above, the tolerance (%) setting can be selected within the range of 20 to 40%. In standard analysis of the L-section of an Al interconnector, it is preferably set to approximately 30%. The steps for adjusting this tolerance are supplementally explained. It is preferable to select or confirm the tolerance value so that the shape and size of the Si phase extracted and identified using the Chi scan function are equivalent to the shape and size of the Si phase identified from the EDS map showing the Si element concentration in a two-dimensional EDS analysis.

[0073] In the present invention, the average value of the shape ratio (c / d) of the Al phase is set to the average value of the values obtained by measuring at least 3 locations. When selecting the measurement area, from the perspective of ensuring the objectivity of the measurement data, it is preferred to obtain a sample for measurement at an interval of 50 cm or more from the Al connecting material to be measured in the direction of the central axis of the Al connecting material, and provide it for measurement. In addition, in the present invention, with respect to the measurement area of the crystal orientation using the EBSD method, it is preferred that the length of the central axis direction of the Al connecting material is 300 μm or more and less than 800 μm, and the entire Al connecting material is inserted in the direction perpendicular to the central axis of the Al connecting material. However, if the size is large and it is difficult to measure the entirety, it can also be adjusted within a range of less than 600 μm.

[0074] - Shape of Si phase in L-section -

[0075] It was found that by setting the average value of the ratio of the short side length e to the long side length f of the Si phase in the L cross section (e / f) to be in the range of greater than 0.2 and less than 0.7, the generation of internal cracks during manufacturing can be suppressed, and good temperature cycle reliability can also be obtained in high-speed TCT.

[0076] The value of this ratio (e / f) is an indicator of flatness. Figure 3 To explain further. Figure 3 This is a diagram schematically showing the Si phase in the L cross section of the Al connecting material, showing that the central axis direction of the Al connecting material corresponds to Figure 3 The horizontal direction (left and right direction) of the central axis corresponds to Figure 3 For the Si phase in the L section, the above “short side length e” is equivalent to Figure 3 The size indicated by the symbol e. In addition, for the Si phase in the L section, the above-mentioned "long side length f" is equivalent to Figure 3 The dimension indicated by the symbol f in the figure. Hereinafter, the ratio (e / f) of the short side length e to the long side length f of the Si phase in the L-section will also be referred to as the "Si phase shape ratio (e / f)." The value of the Si phase shape ratio (e / f) can be determined using the grain shape aspect ratio (Grain Shape Aspect Ratio) in the analysis software included with the device.

[0077] Furthermore, in order to suppress the generation of internal cracks during manufacturing and obtain good temperature cycling reliability even in high-speed TCT, the average value of the aspect ratio (e / f) of the Si phase in the L-section only needs to be within the above-mentioned appropriate range, and it is not necessary for all Si phases to have an aspect ratio (e / f) within the range of 0.2 to 0.7. For example, Si phases with an aspect ratio (e / f) of less than 0.2 may be included, and Si phases with an aspect ratio (e / f) exceeding 0.7 may also be included.

[0078] From the viewpoint of being able to further suppress the generation of internal cracks during manufacturing and obtaining further good temperature cycle reliability in high-speed TCT, the average value of the shape ratio (e / f) of the Si phase in the L-section of the Al connecting material of the present invention is more preferably greater than 0.25, further preferably greater than 0.26, greater than 0.28 or greater than 0.3, and the upper limit is more preferably less than 0.65, further preferably less than 0.64, less than 0.62 or less than 0.6.

[0079] In determining the average value of the shape ratio (e / f) of the Si phase in the L-section of the Al interconnector, a method combining information on the Al concentration and Si concentration obtained by SEM-EDS with information on the crystal orientation obtained by EBSD can be used, similar to the determination of the average value of the shape ratio (c / d) of the Al phase. Therefore, in one embodiment, the average value of the shape ratio (e / f) of the Si phase in the L-section of the Al interconnector of the present invention is calculated by the following steps (1) to (3).

[0080] (1) Using the L-section of the Al interconnector as the inspection surface, the concentrations of Al and Si were measured using EDS and the crystal orientation was measured using EBSD simultaneously.

[0081] (2) Separate and extract Al and Si using the Chi Scan function. Specifically, by setting a tolerance corresponding to the Si threshold based on the Si EDS measurement results, Al and Si can be separated and identified. Crystallographic information on Al and Si from the material file can be used to analyze crystal orientation.

[0082] (3) For the region determined to be the Si phase, the crystal orientation is analyzed. If the orientation difference between the measurement points is 15° or more, it is determined to be a grain boundary and the ratio (e / f) of each grain is calculated. The ratio (e / f) of each grain is averaged to calculate the average value of the shape ratio (e / f) of the Si phase. Here, the average value of the shape ratio (e / f) of the Si phase is calculated in the same way as the average value of the shape ratio (c / d) of the Al phase. The value (grain shape aspect ratio) of the grain shape aspect ratio (Grain Shape Aspect Ratio) of the analysis software is used. In addition, as for the average calculation, the average value obtained by averaging the region that can be selected by the software attached to the device is used, as in the case of calculating the average value of the shape ratio (c / d) of the Al phase. By using the average value obtained by averaging the region, it is possible to accurately measure and determine whether the conditions related to the average value of the shape ratio (e / f) of the Si phase that are suitable for suppressing the occurrence of internal cracks during manufacturing and obtaining good temperature cycle reliability in high-speed TCT are met.

[0083] When measuring the average value of the shape ratio (e / f) of the Si phase in the L cross section, the tolerance setting range in the above step (2), the method of obtaining the sample for measurement, and the measurement area of the crystal orientation using the EBSD method are the same as those previously described for the measurement of the average value of the shape ratio (c / d) of the Al phase.

[0084] - Average diameter of Si phase in L cross section -

[0085] The Al connecting material of the present invention preferably has an average diameter of the Si phase in the L-section of 0.8 μm or more and 4 μm or less.

[0086] Al connecting materials, which have been strengthened by adding Si or other materials, are prone to damaging the semiconductor chip during the first bonding. Adjusting the ultrasonic vibration or load to reduce this damage can sometimes cause a region of insufficient metal bonding near the center of the bonding area between the Al connecting material and the electrode (hereinafter also referred to as a "center gap"). This center gap, due to insufficient metal bonding, can lead to reduced bonding strength and failures in temperature cycling tests.

[0087] If the average diameter of the Si phase in the L-section is within the range of 0.8 μm to 4 μm, central gaps in the first joint can be suppressed. Controlling the average diameter of the Si phase is believed to promote deformation of the Al phase, which contributes to joining, and improve the efficiency of ultrasonic vibration transmission to the central portion of the joining region, thereby suppressing central gaps.

[0088] Furthermore, by combining this with controlling the shape of the Si phase in the L-section to optimize the average diameter of the Si phase, it is effective in further improving the temperature cycling reliability in high-speed TCT. As an example of this effect, in high-speed TCT at -50°C to 175°C, the effect of suppressing the decrease in bonding strength was confirmed even after 12,000 cycles. It is believed that by simultaneously controlling the average diameter and shape of the Si phase, it is possible to improve the effect of reducing the thermal strain of the joint and the effect of suppressing the progression of cracks near the joint interface.

[0089] From the viewpoint of being able to further suppress the intermediate defect of the first joint and further improve the temperature cycle reliability in high-speed TCT, the average diameter of Si in the L cross section of the Al connecting material of the present invention is preferably 3.5 μm or less, more preferably 3.4 μm or less, 3.2 μm or less or 3 μm, and its lower limit is more preferably 1 μm or more, and more preferably 1.2 μm or more.

[0090] The method for measuring the average diameter of the Si phase in the L section of the Al connecting material is described. In the determination of the average diameter of the Si phase in the L section, similarly to the determination of the average value of the shape ratio (c / d) of the Al phase and the average value of the shape ratio (e / f) of the Si phase, a method of combining the information of the Al concentration and Si concentration obtained by SEM-EDS with the information of the crystal orientation obtained by EBSD can be used. The detailed steps can be the same as the steps previously described for the determination of the average value of the shape ratio (e / f) of the Si phase, that is, for the area determined to be the Si phase, the crystal orientation can be analyzed by utilizing the analysis software attached to the device. If the orientation difference between the measurement points is 15° or more, it is judged to be a grain boundary and the equivalent circle diameter is calculated. The average value of the equivalent circle diameter of each Si phase is defined as the average diameter of the Si phase. In the process of obtaining the equivalent circle diameter of the Si phase, the portion where the crystal orientation cannot be measured or the portion where the orientation analysis can be measured but the reliability is low is excluded and calculated. Therefore, in one embodiment, the average diameter of the Si phase in the L section of the Al connecting material of the present invention is calculated by the following steps (1) to (3).

[0091] (1) Using the L-section of the Al interconnector as the inspection surface, the concentrations of Al and Si were measured using EDS and the crystal orientation was measured using EBSD simultaneously.

[0092] (2) Separate and extract Al and Si using the Chi Scan function. Specifically, by setting a tolerance corresponding to the Si threshold based on the Si EDS measurement results, Al and Si can be separated and identified. Crystallographic information on Al and Si from the material file can be used to analyze crystal orientation.

[0093] (3) For the area determined to be the Si phase, the crystal orientation is analyzed. If the orientation difference between the measurement points is 15° or more, it is judged to be a grain boundary and the equivalent circle diameter of each grain is calculated. Then, the equivalent circle diameter of each grain is averaged to calculate the average diameter of the Si phase. Here, for the average calculation, the average value obtained by averaging the regions that can be selected by the software attached to the device is used. By using the average value obtained by averaging the regions, it is possible to accurately measure and determine whether the conditions related to the average diameter of the Si phase that are suitable for suppressing the middle defect of the first joint and further improving the temperature cycle reliability in the high-speed TCT are met.

[0094] In the present invention, when calculating the average diameter of the Si phase in the L-section, only Si phases with a diameter (equivalent circle diameter) of 0.5 μm or greater are considered. This allows for highly accurate determination of whether the requirements for the average diameter of the Si phase in the L-section, which are suitable for suppressing gaps in the first joint and further improving temperature cycling reliability in high-speed TCT, are met.

[0095] When measuring the average diameter of the Si phase in the L cross section, the tolerance setting range in the above step (2), the method of obtaining the sample for measurement, and the measurement area of the crystal orientation using the EBSD method are the same as those previously described for the measurement of the average value of the shape ratio (c / d) of the Al phase.

[0096] - Average diameter of the Al phase in the L cross section -

[0097] The Al connecting material of the present invention preferably has an average diameter of the Al phase in the L-section of 5 μm or more and 40 μm or less.

[0098] In the case where the object of the second joining is a Cu substrate or other material that is harder than Al, in addition, the ultrasonic vibration and load conditions during the second joining are set higher than those during the first joining, and the deformation of the Al connecting material during the second joining, which is strengthened by adding Si, etc., tends to become unstable. In this regard, the inventors found that by making the average diameter of the Al phase in the L cross section within the range of 5 μm to 40 μm, the deviation of the joining strength in the second joining can be reduced. It is believed that the reason is that the following effects work synergistically: by containing a specified concentration of Si and controlling the average value of the shape ratio (c / d) of the Al phase and the average value of the shape ratio (e / f) of the Si phase within a specified range, the deformation of the Al connecting material caused by ultrasonic vibration is promoted, and by setting the average diameter of the Al phase to 5 μm to 40 μm, the deformation of the Al connecting material in both directions parallel and perpendicular to the central axis of the Al connecting material is uniformized.

[0099] From the perspective of further reducing the variation in the bonding strength during the second bond and achieving greater bonding strength stability, the average diameter of the Al phase in the L-section of the Al connecting material of the present invention is more preferably 10 μm or greater, further preferably 12 μm or greater, 14 μm or greater, or 15 μm or greater. Furthermore, the upper limit of the average diameter of the Al phase in the L-section is more preferably 35 μm or less, more preferably 30 μm or less, further preferably 28 μm or less, 26 μm or less, or 25 μm or less.

[0100] In measuring the average diameter of the Al phase in the L-section of the Al connecting material, similar to the measurement of the average diameter of the Si phase, a method can be used that combines the information on the Al concentration and Si concentration obtained by SEM-EDS with the information on the crystal orientation obtained by EBSD. Therefore, in one embodiment, the average diameter of the Al phase in the L-section of the Al connecting material of the present invention is calculated by the following steps (1) to (3).

[0101] (1) Using the L-section of the Al interconnector as the inspection surface, the concentrations of Al and Si were measured using EDS and the crystal orientation was measured using EBSD simultaneously.

[0102] (2) Separate and extract Al and Si using the Chi Scan function. Specifically, by setting a tolerance corresponding to the Si threshold based on the Si EDS measurement results, Al and Si can be separated and identified. Crystallographic information on Al and Si from the material file can be used to analyze crystal orientation.

[0103] (3) For the region identified as the Al phase, the crystal orientation is analyzed. If the orientation difference between the measurement points is 15° or more, it is determined to be a grain boundary and the equivalent circle diameter of each grain is calculated. Then, the equivalent circle diameters of each grain are averaged to calculate the average diameter of the Al phase. Here, regarding the average calculation, similar to the measurement of the average diameter of the Si phase, the average value calculated based on the regional average that can be selected by the device's attached software is used. By using the average value calculated based on the regional average, it is possible to accurately measure and determine whether the conditions related to the average diameter of the Al phase suitable for improving the stability of the bonding strength in the second bonding are met.

[0104] In the present invention, when calculating the average diameter of the Al phase in the L-section, only Al phases with a diameter (equivalent circle diameter) of 0.5 μm or greater are considered. This allows for highly accurate determination of whether the requirements for the average diameter of the Al phase in the L-section, which are suitable for improving the stability of the bonding strength in the second bonding process, are met.

[0105] When measuring the average diameter of the Al phase in the L cross section, the tolerance setting range in the above step (2), the method of obtaining the sample for measurement, and the measurement area of the crystal orientation using the EBSD method are the same as those described above for the measurement of the average value of the shape ratio (c / d) of the Al phase.

[0106] In addition, as for the method of measuring the shape ratio (c / d) of the Al phase, the shape ratio (e / f) of the Si phase, the average diameter of the Al phase, and the average diameter of the Si phase, in addition to the above-mentioned methods, there are several methods including binarization processing based on the observation image of the L cross section. However, in the present invention, since it has multiple measurement functions, multiple characteristics such as the shape ratio (c / d) of the Al phase, the shape ratio (e / f) of the Si phase, the average diameter of the Al phase, and the average diameter of the Si phase can be obtained in a single measurement, and automatic analysis can be performed. As described above, a method of combining the information on the Al concentration and Si concentration obtained by SEM-EDS with the information on the crystal orientation obtained by EBSD is used because it is equipped with multiple measurement functions and can perform automatic analysis. It is a popular device and analysis technology and the measurement is easy.

[0107] - Addition of Sr, Ca, Na, Ti, and P -

[0108] The Al connecting material of the present invention may further contain a total of 10 mass ppm to 800 mass ppm of any one or more of Sr, Ca, Na, Ti, and P (hereinafter also referred to as "first element group").

[0109] By also containing a total of more than 10 mass ppm and less than 800 mass ppm of any one of Sr, Ca, Na, Ti, and P, the frequency of wire breakage during the wire drawing process of the Al connecting material can be reduced. Al alloys containing Si at a high concentration of 3.0 mass % and less than 12.0 mass % tend to have an increased frequency of wire breakage during the wire drawing process. It is believed that one of the reasons is that the particles of the Si phase crystallized during solidification cause stress concentration during the wire drawing process, inducing wire breakage. It is speculated that by adding the first element group, the stress concentration in the wire drawing can be alleviated and wire breakage can be reduced by utilizing the effects of being able to evenly distribute the granular Si phase and being able to inhibit the growth and coarsening of the Si phase. It is believed that by controlling the shapes of the Al phase and the Si phase in the L cross section and adding the first element group, the effect of alleviating stress concentration in the wire drawing can be improved.

[0110] From the viewpoint of reducing the frequency of wire breakage during wire drawing, the total concentration of the first element group in the Al connecting material of the present invention is more preferably 20 mass ppm or more, further preferably 30 mass ppm or more, 40 mass ppm or more or 50 mass ppm or more, and the upper limit is preferably 750 mass ppm or less, further preferably 740 mass ppm or less, 720 mass ppm or less or 700 mass ppm or less.

[0111] - Addition of Mn, Sn, Cu, and Zn -

[0112] The Al connecting material of the present invention may further contain a total of 10 mass ppm to 500 mass ppm of any one or more of Mn, Sn, Cu, and Zn (hereinafter also referred to as "second element group").

[0113] By also containing any one or more of Mn, Sn, Cu, and Zn totaling to more than 10 mass ppm and less than 500 mass ppm, it is possible to suppress the damage on the surface of the Al connecting material, the generation of scraping, and form a smooth surface. The Al alloy containing Si at a high concentration of more than 3.0 mass % and less than 12.0 mass %, due to the hardening of the surface, the shedding of Si phase and Al oxide present on the surface, etc., thereby there is damage and scraping on the surface during wire drawing, and belongs to the situation of the Al connecting material with large surface unevenness. It is speculated that by adding the second element group, the stabilization of the Al oxide on the surface of the Al connecting material, the friction between the Al connecting material and the mold, etc., can be reduced, thereby reducing the damage and scraping in the wire drawing. It is believed that by controlling the shape of the Al phase and the Si phase respectively in the L cross section and adding the second element group, it is possible to suppress the damage on the surface of the Al connecting material, the generation of scraping and improve the effect of forming a smooth surface.

[0114] From the viewpoint of suppressing surface damage and scraping to form an Al connecting material with a smooth surface, the total concentration of the second element group in the Al connecting material of the present invention is more preferably 20 mass ppm or more, further preferably 30 mass ppm or more, 40 mass ppm or more or 50 mass ppm or more. From the viewpoint of suppressing damage to the semiconductor chip and easily achieving good first bonding strength, the upper limit is more preferably 450 mass ppm or less, further preferably 440 mass ppm or less, 420 mass ppm or less or 400 mass ppm or less.

[0115] As the aluminum raw material for producing the Al connecting material of the present invention, Al with a purity of 4N (Al: 99.99 mass % or more) is preferably used, and 5N (Al: 99.999 mass % or more) Al with a low impurity content is more preferably used.

[0116] In the range that does not hinder the effect of the present invention, the Al connecting material of the present invention may also contain elements (hereinafter, also referred to as "other elements") other than the first element group and the second element group. The total concentration of the other elements in the Al connecting material is not particularly limited in the range that does not hinder the effect of the present invention. The total concentration of the other elements may, for example, be less than 0.5 mass %, less than 0.4 mass %, less than 0.3 mass %, less than 0.2 mass %, less than 0.15 mass %, less than 0.1 mass %, less than 0.08 mass %, less than 0.06 mass %, less than 0.05 mass %, less than 0.04 mass %, less than 0.03 mass %, less than 0.02 mass %, less than 0.018 mass %, less than 0.016 mass %, less than 0.015 mass %, less than 0.014 mass %, less than 0.012 mass % or less than 0.01 mass %. The lower limit of the total concentration of other elements is not particularly limited and may be 0 mass %.

[0117] In one embodiment, the remainder of the Al connecting material of the present invention consists of Al and inevitable impurities. Therefore, in a preferred embodiment, the Al connecting material of the present invention consists of Al, Si and inevitable impurities. In another preferred embodiment, the Al connecting material of the present invention consists of Al, Si, any one or more of the first element group, and inevitable impurities. Furthermore, in another preferred embodiment, the Al connecting material of the present invention consists of Al, Si, any one or more of the second element group, and inevitable impurities. Furthermore, in another preferred embodiment, the Al connecting material of the present invention consists of Al, Si, any one or more of the first element group, any one or more of the second element group, and inevitable impurities.

[0118] In a preferred embodiment, the Al connecting material of the present invention does not have a coating mainly composed of a metal other than Al on the periphery of the Al connecting material. Here, "coating mainly composed of a metal other than Al" means a coating containing a metal other than Al of 50% by mass or more.

[0119] The Al connecting material of the present invention may be either an Al bonding wire or an Al bonding tape. When the Al connecting material of the present invention is an Al bonding wire, its wire diameter is not particularly limited and may, for example, be in the range of 100 to 600 μm. When the Al connecting material of the present invention is an Al bonding tape, the dimensions (W×T) of its rectangular or substantially rectangular cross-section are not particularly limited and may, for example, be 100 to 3000 μm and 50 to 600 μm.

[0120] The Al connecting material of the present invention can suppress the generation of internal cracks during manufacturing and can provide excellent temperature reliability. Therefore, the Al connecting material of the present invention can be suitably used as an Al connecting material for semiconductor devices, especially an Al connecting material for power semiconductor devices.

[0121] - Method for manufacturing Al connecting material -

[0122] An example of a method for producing the Al connecting material of the present invention will be described below. An example of producing an Al bonding wire having a wire diameter of 200 to 400 μm will be described.

[0123] The higher the purity of Al and alloying elements used as raw materials, the more preferred it is. Al preferably has a purity of 99.99% by mass or more, with the remainder consisting of inevitable impurities. Si, the first element group, and the second element group used as alloying elements preferably have a purity of 99.9% by mass or more, with the remainder consisting of inevitable impurities. The Al alloy used for bonding wire can be produced by loading Al raw materials and alloying element raw materials into a crucible made of graphite or alumina processed in a manner to obtain a cylindrical ingot, and melting them using an electric furnace or a high-frequency heating furnace. The diameter of the cylindrical ingot is preferably Φ6 mm or more and less than Φ8 mm in consideration of processability in subsequent processing steps. Regarding the atmosphere in the furnace during melting, in order to prevent excessive oxidation of Al and other elements constituting the wire, an inert atmosphere or a reducing atmosphere is preferably used. Regarding the maximum temperature reached by the melt during melting, considering ensuring the fluidity of the melt and facilitating control of the shape and size of the Si phase during solidification, it is preferably in the range of 800°C or more and less than 1050°C. The cooling method after melting can be water cooling, furnace cooling, air cooling, etc.

[0124] The cylindrical ingot obtained by melting is subjected to solution treatment by heating at high temperature, and then repeatedly subjected to wire drawing using a die to produce a wire of the desired diameter. The wire after wire drawing is subjected to final heat treatment in an electric furnace and can be used as an Al alloy bonding wire.

[0125] To control the shape of the Al and Si phases in the L-section, it is effective to control the heat treatment conditions such as solution treatment and final heat treatment, as well as the wire drawing conditions. During wire drawing, it is effective to use a lubricant to ensure lubricity at the contact interface between the wire and the die.

[0126] Below is an example of manufacturing conditions for controlling the average value of the shape ratio (c / d) of the Al phase in the L-section to be within the range of not less than 0.25 and not more than 0.7, and for controlling the average value of the shape ratio (e / f) of the Si phase in the L-section to be within the range of not less than 0.2 and not more than 0.7.

[0127] In order to adjust the shape of the Si phase (shape ratio (e / f)), it is effective to control the linear area reduction rate during solution treatment and wire drawing.

[0128] The temperature range for solution treatment of the ingot is preferably 400°C or higher and less than 550°C, and the duration is preferably 1 hour or higher and less than 6 hours. This solution treatment promotes the severing and spheroidization of the Si phase that crystallizes during solidification, thereby controlling the shape of the Si phase during subsequent wire drawing. For example, when solution treatment is performed at high temperatures, the shape ratio (e / f) of the Si phase tends to increase.

[0129] Regarding the wire drawing conditions, it is effective to set the linear area reduction rate per die used in the wire drawing process to be 20% or more and less than 35%. Here, if the linear area reduction rate per die is defined as P1, P1 is expressed by the following formula.

[0130] P1={(R2 2 -R1 2 ) / R2 2}×100

[0131] Here, R2 represents the diameter (mm) of the wire before processing, and R1 represents the diameter (mm) of the wire after processing.

[0132] By adjusting the wire area reduction rate to the higher range (high wire area reduction rate) compared to conventional wire drawing conditions, the entire wire is significantly deformed during die processing, increasing the processing strain within the wire, aligning the Si phase along the wire's central axis while adjusting its shape. By combining control of the solution treatment conditions described above with control of the wire area reduction rate during wire drawing, the shape ratio (e / f) of the Si phase can be easily adjusted.

[0133] In order to adjust the shape (shape ratio (c / d)) of the Al phase in the L-section, it is effective to control the wire feed speed during the wire drawing process and the final heat treatment.

[0134] The wire feeding speed during wire drawing will be described. The wire drawing process from the ingot (wire diameter D0) obtained by melting to a wire diameter of about half the final wire diameter (wire diameter D) (intermediate wire diameter: D+0.5(D0-D)) is referred to as "wire drawing process 1", and the wire drawing process from the intermediate wire diameter to the final wire diameter is referred to as "wire drawing process 2". It is effective to set the wire feeding speed in wire drawing process 1 to be greater than 5 m / min and less than 15 m / min, and to set the wire feeding speed in wire drawing process 2 to be greater than 20 m / min and less than 50 m / min. By setting the wire feeding speed to the above range, it is possible to combine the control of processing strain and dynamic recrystallization during wire drawing, and it is easy to control the shape ratio (c / d) of the Al phase in the L cross section of the wire to be within the target range. As needed, by combining the adjustment of the above-mentioned wire area reduction rate to a range of greater than 20% and less than 35% with the adjustment of the wire feeding speed, the shape ratio (c / d) of the Al phase can be controlled with high precision.

[0135] Effective conditions for the final heat treatment are a temperature range of 250°C to less than 350°C and a duration of 2 hours to less than 24 hours. Using the strain within the wire produced by the wire drawing process as a driving force, the recrystallization of the Al phase can be promoted, adjusting the shape of the grains. For example, when the final heat treatment is performed at a low temperature for a long time, the Al phase grains become granular, and the shape ratio (c / d) tends to increase.

[0136] In order to adjust the average diameter of the Si phase in the L cross section to a range of 0.8 μm to 4 μm, it is effective to adjust the temperature during melting within a range of 800°C to less than 1050°C, adjust the casting temperature within a range of 700°C to less than 780°C, control the temperature of the solution treatment to be 450°C to 550°C, and control the time to be 1 hour to 6 hours. The so-called casting temperature refers to the temperature when the molten melt is cast into a mold, etc., which is equivalent to the solidification start temperature. If the casting temperature is high, the Si phase crystallized during solidification tends to coarsen and columnarize, and the average diameter of the Si phase tends to increase. If the temperature of the solution treatment is high, the columnar Si phase is cut and granulated, and the average diameter of the Si phase tends to decrease. In order to further reduce the average diameter of the Si phase, it is effective to increase the cooling rate during solidification, for example, water cooling is also effective.

[0137] In order to control the average diameter of the Al phase in the L-section to a range of 5 μm to 40 μm, it is effective to adjust the temperature and time of the heat treatment at the final wire diameter to control the growth of grains caused by the recrystallization of the Al phase. It is effective to control the temperature range of the final heat treatment to a range of 250°C to less than 340°C and the time to a range of 5 hours to less than 24 hours. By performing a homogenization treatment at a relatively low temperature and adjusting the amount of Si dissolved in the Al phase, it is easy to adjust the recrystallization temperature of the Al phase during the final heat treatment, and the size of the Al phase can be controlled to the target range.

[0138] In addition, if an intermediate heat treatment is carried out as needed, it is easy to adjust the conditions of the above-mentioned final heat treatment. The so-called intermediate heat treatment is a heat treatment carried out in the middle of the process from ingot processing to the final wire diameter. It is effective to set the temperature range of the intermediate heat treatment to above 250°C and less than 400°C, and to set the time to above 30 minutes and less than 3 hours. It is effective to carry out the intermediate heat treatment at a wire diameter of 2.5 to 4.0 times the final wire diameter. By carrying out the intermediate heat treatment, the reduction of the processing strain of the Al phase and the progress of recrystallization can be achieved. By adjusting the Al phase once, it becomes easy to adjust the particle size of the Al phase in the subsequent final heat treatment. For example, if the intermediate heat treatment temperature is increased, the diameter of the Al phase tends to decrease.

[0139] As mentioned above, the above description describes an example of the manufacture of an Al bonding wire as a wire rod as a representative example of an Al connecting material. Al bonding strips as strip materials can basically be manufactured using the same steps. The temperature and time of the heat treatment can be roughly the same as those described above. Furthermore, when manufacturing an Al bonding strip by rolling, it is sufficient to replace the die area reduction rate with the reduction rate and adjust the process.

[0140] [Semiconductor devices]

[0141] By using the Al connecting material of the present invention to connect electrodes on a semiconductor chip to external electrodes on a lead frame or substrate, a semiconductor device can be manufactured. As described above, both the first bonding to the electrodes on the semiconductor chip and the second bonding to the electrodes on the lead frame or substrate use wedge bonding.

[0142] In one embodiment, a semiconductor device of the present invention includes a circuit substrate, a semiconductor chip, and an Al connecting material for electrically connecting the circuit substrate and the semiconductor chip, wherein the Al connecting material is the Al connecting material of the present invention.

[0143] In the semiconductor device of the present invention, the circuit substrate and the semiconductor chip are not particularly limited, and known circuit substrates and semiconductor chips that can be used to constitute a semiconductor device can be used. Alternatively, a lead frame can be used instead of the circuit substrate. For example, as in the semiconductor device described in Japanese Patent Application Laid-Open No. 2020-150116, a semiconductor device structure including a lead frame and a semiconductor chip mounted on the lead frame can also be used.

[0144] As semiconductor devices, there are various semiconductor devices used in electrical products (such as computers, mobile phones, digital cameras, televisions, air conditioners, solar power generation systems, etc.) and transportation vehicles (such as motorcycles, cars, trams, ships and aircraft, etc.), among which semiconductor devices for power (power semiconductor devices) are preferred.

[0145] [Example]

[0146] Hereinafter, the present invention will be described in detail with reference to Examples, but the present invention is not limited to the following Examples.

[0147] (sample)

[0148] The method for preparing the sample is described. As the raw material, Al having a purity of 4N (99.99% by mass or more) and the remainder consisting of inevitable impurities is used. As the alloying elements, Si, the first element group (Sr, Ca, Na, Ti, P), and the second element group (Mn, Sn, Cu, Zn) are alloying elements having a purity of 99.99% by mass or more and the remainder consisting of inevitable impurities. The Al alloy used for the Al connecting material is manufactured by filling an alumina crucible with Al raw materials and raw materials of alloying elements and melting them in a high-frequency heating furnace. The atmosphere in the furnace during melting is an Ar atmosphere, the maximum temperature reached by the melt during melting is 800°C or more and less than 1050°C, and the casting temperature is 700°C or more and less than 780°C. The cooling method after melting is air cooling in the atmosphere or water cooling in water.

[0149] A cylindrical ingot of Φ6 mm is obtained by melting. After the ingot is subjected to solution treatment and homogenization treatment, it is subjected to wire drawing using a mold and intermediate heat treatment to produce an Al connecting material (Al bonding wire) of Φ300 μm. The temperature range of the solution treatment is above 450°C and below 550°C, and the time is above 1 hour and below 4 hours. After the solution treatment is completed, homogenization treatment is continuously carried out in the middle of cooling. The temperature range of the homogenization treatment is above 250°C and below 350°C, and the time is above 2 hours and below 6 hours. The cooling method after the homogenization treatment is air cooling in the atmosphere.

[0150] A commercially available lubricant is used during the wire drawing process, and the wire area reduction rate per die during the wire drawing process is 25% or more and less than 35%. The final heat treatment temperature range is 250°C or more and less than 350°C, and the final heat treatment time is 5 hours or more and less than 24 hours.

[0151] In some embodiments, the wire feed speed during wire drawing process 1 is set to 10 m / min or higher and less than 15 m / min, and the wire feed speed during wire drawing process 2 is set to 20 m / min or higher and less than 40 m / min. In some embodiments, the intermediate heat treatment temperature range is 250°C or higher and less than 350°C, and the duration is 30 minutes or higher and less than 3 hours. The intermediate heat treatment is performed once, at a wire diameter of 2.5 to 4.0 times the final wire diameter.

[0152] (Method for determining element content)

[0153] Concentration analysis of elements contained in the Al connecting material was performed using ICP-OES ("PS 3520UVDDII" manufactured by Hitachi High-Tech Science Co., Ltd.) or ICP-MS ("Agilent 7700x ICP-MS" manufactured by Agilent Technologies, Ltd.) as an analysis apparatus.

[0154] (Method for measuring the shape of Al phase and Si)

[0155] The L cross section (cross section in the direction of the central axis including the central axis) of the Al connecting material is used as the inspection surface to measure the shapes of the Al phase and the Si phase (ratio (c / d), ratio (e / f)). In the present invention, the central axis of the Al connecting material and the cross section in the direction of the central axis including the central axis (L cross section) are as follows: Figure 1 shown. Figure 1 While the Al connecting material is shown as an Al bonding wire with a circular cross-section, when the Al connecting material is an Al bonding tape with a rectangular or substantially rectangular cross-section of width W and thickness T, the central axis refers to the axis passing through the center of width W and the center of thickness T. Furthermore, the L-section refers to a cross-section in the direction of the central axis that includes the central axis and is a cross-section in the direction of thickness T. When cross-sectioning is performed to expose the L-section of the Al connecting material, the cross-section may deviate from the central axis of the Al connecting material. In this case, if the length of the L-section in the direction perpendicular to the central axis is at least 90% of the wire diameter of the Al connecting material (thickness T in the case of a tape), the cross-section is considered to include the central axis.

[0156] In addition, FE-SEM (SU-70 manufactured by Hitachi High-Tech Corporation) is used in the measurement, and APEX (for data collection), OIM Data Collection (for Chi scanning), and OIM Anaysis (data analysis) manufactured by TSLSolutions are used as the analysis software. Three measurement areas are randomly selected at intervals of 50 cm or more in the central axis direction of the Al connecting material, and the three areas are measured. The measurement area is determined in a manner that is 300 μm or more and less than 800 μm in the central axis direction of the Al connecting material, and the Al connecting material enters as a whole in a direction perpendicular to the central axis. In addition, the main conditions for EDS and EBSD measurements are an acceleration voltage of 15 kV, a scanning speed of 30 to 120 points / second, a measurement magnification of 350 times, and a measurement interval of 0.1 to 0.3 μm. Here, when the scanning speed is fast, the measurement time can be shortened, but there is a concern that the measurement accuracy of EDS will be reduced. It is preferred to select an appropriate scanning speed within the above range.

[0157] - Shape of Al phase -

[0158] The shape of the Al phase (ratio c / d) in the L-section of the Al interconnector was determined by combining information on Al and Si concentrations obtained by SEM-EDS with information on crystal orientation obtained by EBSD. Specifically, the measurement was performed according to the following steps (1) to (3).

[0159] (1) In a measurement area using the L-section of the Al interconnector as an inspection surface, the concentrations of Al and Si were measured using EDS and the crystal orientation was measured using EBSD simultaneously.

[0160] (2) Using the Chi-scan function of the EBSD analysis software, Al and Si are separated and extracted. Specifically, based on the EDS measurement results of Si, a tolerance corresponding to the Si threshold is set, enabling separation and identification of Al and Si. Crystallographic information on Al and Si from the material file is used for crystal orientation analysis. Here, the tolerance condition is primarily set to 30%, and is adjusted as needed.

[0161] (3) For the area determined to be the Al phase, the crystal orientation is analyzed. If the orientation difference between the measurement points is 15° or more, it is judged to be a grain boundary and the shape ratio (c / d) of each grain is calculated. Then, the shape ratio (c / d) of each grain is averaged to calculate the average value of the shape ratio (c / d) of the Al phase. Here, the average value of the shape ratio (c / d) of the Al phase is calculated using the value of the grain shape aspect ratio (Grain Shape Aspect Ratio) of the analysis software ("Grain Shape Aspect Ratio"). Here, for the calculation method of the grain shape aspect ratio, the software automatically calculates the ratio (c / d) of the short side length (c) (GrainShape Minor Axis: grain shape short axis) and the long side length (d) (Grain Shape Major Axis: grain shape long axis) of a grain. Here, for the average calculation, the average value obtained by regional averaging is used.

[0162] The average value of the shape ratio (c / d) of the Al phase is the average value of the values obtained by the above-mentioned steps (1) to (3) for the three measurement areas.

[0163] - Shape of Si phase -

[0164] The measurement of the Si phase aspect ratio (e / f) in the L-section of the Al interconnector is similar to the measurement of the Al phase aspect ratio (c / d). This method combines information on Al and Si concentrations obtained by SEM-EDS with information on crystal orientation obtained by EBSD. Specifically, after performing the above steps (1) and (2), the measurement is performed according to the following step (3).

[0165] (3) For the region identified as the Si phase, the crystal orientation is analyzed. If the orientation difference between the measurement points is 15° or more, it is determined to be a grain boundary and the shape ratio (e / f) of each grain is calculated. Then, the shape ratio (e / f) of each grain is averaged to calculate the average value of the shape ratio (e / f) of the Si phase. Here, the average value of the shape ratio (e / f) of the Si phase is determined in the same way as the average value of the shape ratio (c / d) of the Al phase. The value of the grain shape aspect ratio (Grain Shape Aspect Ratio) ("Grain Shape Aspect Ratio") of the analysis software is used for the average calculation. For the average calculation, the average value obtained by regional averaging is used.

[0166] The average value of the shape ratio (e / f) of the Si phase is the average value of the values obtained by the above-mentioned steps (1) to (3) for the three measurement areas.

[0167] (Method for measuring average diameters of Al phase and Si phase)

[0168] The average diameters of the Al phase and Si phase in the L cross section of the Al connecting material are measured using a method that combines the information on the Al concentration and Si concentration obtained by SEM-EDS with the information on the crystal orientation obtained by EBSD, similar to the measurement of the shapes of the Al phase and Si phase.

[0169] - Average diameter of Al phase -

[0170] Specifically, after the above steps (1) and (2) are performed, the measurement is performed according to the following step (3).

[0171] (3) For the region identified as the Al phase, the crystal orientation is analyzed. If the orientation difference between the measurement points is 15° or more, it is determined to be a grain boundary and the equivalent circle diameter of each grain is calculated. Then, the equivalent circle diameter of each grain is averaged to calculate the average diameter of the Al phase. In this average calculation, the average value obtained by regional averaging is used. In addition, when calculating the average diameter of the Al phase in the L-section, only Al phases with a diameter (equivalent circle diameter) of 0.5 μm or more are included.

[0172] The average diameter of the Al phase is the average value of the values obtained by the above-mentioned steps (1) to (3) for the three measurement areas.

[0173] - Average diameter of Si phase -

[0174] Specifically, after the above steps (1) and (2) are performed, the measurement is performed according to the following step (3).

[0175] (3) For the region identified as the Si phase, the crystal orientation is analyzed. If the orientation difference between the measurement points is 15° or more, it is determined to be a grain boundary and the equivalent circle diameter of each grain is calculated. Then, the equivalent circle diameter of each grain is averaged to calculate the average diameter of the Si phase. Here, in the average calculation, the average value obtained by regional averaging is used. In addition, when calculating the average diameter of the Si phase in the L cross section, only the Si phase with a diameter (equivalent circle diameter) of 0.5 μm or more is used as the object.

[0176] The average diameter of the Si phase is the average value of the values obtained by the above-mentioned steps (1) to (3) for three measurement areas.

[0177] (Evaluation Method of Al Connecting Materials)

[0178] The evaluation method of the A1 connecting material is described below. The wire diameter of the A1 connecting material (A1 bonding wire) used in the evaluation is Φ300μm. A semiconductor chip made of Si is used, and an electrode on the semiconductor chip is used, in which an alloy composed of Al-0.5% Cu is formed into a film with a thickness of 4μm. A substrate is used in which a 15μm Ni film is formed on an Al alloy. A commercially available wire bonder (manufactured by Ultrasonic Industry Co., Ltd.) is used for bonding the Al connecting material, and both the first bond and the second bond are wedge bonds.

[0179] (Evaluation method for high-speed temperature cycle reliability)

[0180] A commercially available high-speed thermal shock tester is used in the evaluation of the high-speed temperature cycle test (high-speed TCT). In the high-speed TCT, hot air is blown onto the sample for rapid heating. The sample subjected to the high-speed TCT is a structure in which a semiconductor chip is mounted on a substrate, and the electrodes on the semiconductor chip and the electrodes on the substrate are connected with an Al connecting material. For the sample set in the sample chamber of the high-speed thermal shock tester, heating and cooling are regarded as one cycle, and heat load is repeatedly applied. The lowest temperature during cooling is -50°C, and the highest temperature during heating is 175°C. The heating time including the heating time is 20 seconds, and the cooling time including the cooling time is 40 seconds. After the start of the test, the sample is taken out after 10,000 cycles and 12,000 cycles, and a shear strength test of the first joint is performed. For the value of the shear strength of the first joint used in the evaluation of the high-speed temperature cycle reliability, the average value of the shear strength of 10 randomly selected first joints is used. The ratio (percentage) of the average shear strength after the implementation of the high-speed TCT to the average shear strength before the test is used as the strength maintenance rate. The higher the strength maintenance rate, the better the reliability of the joint. If the strength retention rate is 85% or higher, it is judged as excellent and indicated by a "3." If it is 75% or higher and less than 85%, it is judged as excellent and indicated by a "2." If it is 70% or higher and less than 75%, it is judged as requiring improvement and indicated by a "1." If it is less than 60%, it is considered to have practical problems and indicated by a "0." "3" and "2" are judged as acceptable, while "1" and "0" are judged as unacceptable. The evaluation results are recorded in the "High-Speed Temperature Cycle Reliability" column in the table. The temperature cycle reliability requirement for next-generation power semiconductor devices is equivalent to 10,000 cycles.

[0181] (Evaluation method for internal cracks)

[0182] The evaluation method of internal cracks of the A1 connecting material is described. The manufactured A1 connecting material is evaluated by observation using a soft X-ray projection inspection device (manufactured by Song Ding Precision, μB2600) (hereinafter referred to as X-ray observation). The measurement conditions of the X-ray observation can be appropriately determined according to the wire diameter of the A1 connecting material, but in the case of the A1 connecting material with a wire diameter of 300μm manufactured in this embodiment, the voltage is adjusted to a range of 50 to 80kV and the current is adjusted to a range of 60 to 90μA. Three places are randomly selected at intervals of more than 1m in the central axis direction of the A1 connecting material, and three samples with a length of about 8cm are selected at each of the three places, for a total of 9 samples as measurement samples. Figure 5An example of X-ray observation of an Al connecting material with a wire diameter of 300μm is shown, in which internal cracks were observed. If the length of the internal crack is 0.3mm or longer, it is judged to be a problematic defect and scored as "2". If it is 0.1mm or longer and less than 0.3mm, it is judged to be a problem and scored as "0.5". The total score of the measured parts is set as the "crack index". For the crack index of the entire measured sample, if it is zero, it is judged to be good and evaluated as "3". If it is in the range of 0.1 to 2.0, it is judged to be practical and evaluated as "2". If it is in the range of 2.0 to 5.0, it is judged to need improvement and evaluated as "1". If it exceeds 6.0, it is judged to be difficult to use and evaluated as "0". The evaluation results are recorded in the "Internal Cracks" column in the table.

[0183] (Evaluation Method for Middle Defect of First Joint)

[0184] The evaluation method of the defective middle defect of the first joint is described. After the shear strength test of the above-mentioned first joint is performed, the indentation of the fracture surface on the electrode side is observed with an optical microscope or SEM, and the portion in the fracture area where no metal bonding is obtained is determined to be a middle defect. The portion where the middle defect occurs is a portion that is not bonded even if the electrode is deformed, and can be distinguished from the area where the joint is metal bonded. The shear strength test is carried out under the above-mentioned conditions, and the fracture surfaces of the first joint are observed at 10 locations. Then, the ratio of the total length (K) of the bonding width direction of the middle defect area to the bonding length (J) in the direction perpendicular to the center axis (bonding width direction) of the Al connecting material is calculated as the middle defect ratio (K / J) ( Figure 4 The center-chip ratio was determined at 10 fracture surfaces, and the maximum value was defined as the "center-chip defect rate." A center-chip yield rate of less than 5% was considered good and scored "3." A yield rate of 5% or more and less than 15% was considered practically acceptable and scored "2." A yield rate of 15% or more and less than 25% was considered necessary for improvement and scored "1." A yield rate exceeding 25% was considered a barrier to mass production and scored "0." The evaluation results are recorded in the "Center-chip yield rate for first joint" column in the table.

[0185] (Evaluation Method of Bonding Strength Stability of Second Bonding Portion)

[0186] Shear strength tests were conducted on 30 randomly selected second joints. The joint strengths were measured and the overall standard deviation (σ) was calculated. A σ of 70 gf or greater was considered to be practically problematic and scored "0." A σ of 50 gf or greater but less than 70 gf was considered good and scored "1." A σ of 30 gf or greater but less than 50 gf was considered excellent and scored "2." A σ of less than 30 gf was considered exceptionally excellent and scored "3." "0" was considered unsatisfactory, while "1," "2," and "3" were considered acceptable. The evaluation results are recorded in the "Joint Strength Stability of Second Joints" column in the table.

[0187] (Evaluation method for wire breakage during processing)

[0188] The evaluation method of wire breakage during processing is explained. To wire diameter Carry out wire drawing and check the number of wire breaks. The conveying speed, wire area reduction rate, etc. as the processing conditions for wire drawing are selected from the above conditions, and appropriate manufacturing conditions are adjusted and changed for each wire. The length of the Al connecting material after wire drawing is in the range of 100 to 200 m, and the number of wire breaks is calculated per 100 m. If the number of wire breaks is 0, it is judged to be good and evaluated as "3". If it is 1, it is judged to be able to cope with it by improving the manufacturing conditions and evaluated as "2". If it is 2 to 4 times, the reduction in productivity is regarded as a problem and evaluated as "1". If it is 5 times or more, it is judged to be difficult to use and evaluated as "0". The evaluation results are recorded in the "Wire Breakage During Processing" column in the table.

[0189] (Evaluation method for surface damage and scratches)

[0190] Regarding the surface properties of the A1 connecting material, attention was paid to damage and scraping. The wire diameter of the A1 connecting material is Three measurement areas were randomly selected at intervals of more than 1 m in the central axis direction of the A1 connecting material, and three samples of about 2 cm in length were collected at each of the three locations, and a total of 9 samples were observed. The surface was observed at a magnification of 50 to 500 times the SEM. Damage of more than 50 μm in length and scraping of more than 30 μm in length were judged as poor. The number of damaged and scraped areas was counted. If there were 0, it was good and judged to be qualified, and evaluated as "3". If there were 2 or fewer, it was judged to be no practical problem and evaluated as "2". If there were 3 to 7, it was judged to be a poor surface shape and evaluated as "1". If there were 8 or more, it was judged to be difficult to use and evaluated as "0". The evaluation results are recorded in the "Surface Properties" column in the table.

[0191] The evaluation results of Examples and Comparative Examples are shown in Tables 1 to 3.

[0192] [Table 1]

[0193]

[0194] [Table 2]

[0195]

[0196] [Table 3]

[0197]

[0198] The Al connecting materials of Examples No. 1 to 51 were all confirmed to contain Si in an amount of not less than 3.0 mass % and not more than 12.0 mass %, and the average value of the shape ratio (c / d) of the Al phase was not less than 0.25 and not more than 0.7, and the average value of the shape ratio (e / f) of the Si phase was not less than 0.2 and not more than 0.7. They were able to suppress the generation of internal cracks during manufacturing and also exhibited good temperature cycling reliability in high-speed TCT.

[0199] In addition, the Al connecting materials of Examples Nos. 1, 3 to 19, 21 to 30, and 32 to 51, in which the average diameter of the Si phase in the L cross section was 0.8 μm or more and 4 μm or less, tended to obtain better temperature cycle reliability in high-speed TCT.

[0200] The Al connecting materials of Examples Nos. 1 to 4, 6 to 20, 22 to 24, 26 to 43, 45, 46, and 48 to 51, in which the average diameter of the Al phase in the L cross section was 5 μm to 40 μm, were found to be more likely to achieve better results in terms of bonding strength stability at the second bonding portion.

[0201] Furthermore, the Al connecting materials of Examples Nos. 29 to 39 and 49 to 51 containing at least one of the first element group (Sr, Ca, Na, Ti, P) in a total amount of 10 mass ppm to 800 mass ppm were confirmed to be able to reduce the frequency of wire breakage during processing.

[0202] The Al connecting materials of Examples Nos. 41 to 46 and 48 to 51 containing at least one of the second element group (Mn, Sn, Cu, Zn) in a total amount of 10 mass ppm to 500 mass ppm were confirmed to suppress surface damage and scraping and have a smooth surface.

[0203] On the other hand, the Si concentration of the Al connecting material, the average value of the shape ratio (c / d) of the Al phase, and the average value of the shape ratio (e / f) of the Si phase in Comparative Examples No. 1 to 7 are all outside the scope of the present invention, and it is confirmed that the effect of suppressing the generation of internal cracks during manufacturing and the high-speed temperature cycle reliability cannot be fully obtained.

Claims

1. An Al connecting material comprising 3.0 mass % or more and 12.0 mass % or less of Si, The average value of the ratio c / d of the short side length c to the long side length d of the Al phase in the L cross section of the Al connecting material, that is, the cross section in the central axis direction including the central axis, is 0.25 or more and 0.7 or less. The average value of the ratio e / f of the short side length e to the long side length f of the Si phase in the L cross section of the Al interconnector is 0.2 or more and 0.7 or less.

2. The Al connecting material according to claim 1, The average diameter of the Si phase in the L cross section is 0.8 μm or more and 4 μm or less.

3. The Al connecting material according to claim 1 or 2, The average diameter of the Al phase in the L cross section is 5 μm or more and 40 μm or less.

4. The Al connecting material according to any one of claims 1 to 3, It further contains at least one of Sr, Ca, Na, Ti, and P in an amount of 10 mass ppm to 800 mass ppm in total.

5. The Al connecting material according to any one of claims 1 to 4, It further contains at least one of Mn, Sn, Cu, and Zn in a total amount of 10 mass ppm to 500 mass ppm.

Images