copper-ceramic composite and a module containing this composite

Abstract

copper-ceramic composite, comprising - a ceramic substrate, - a coating of copper or a copper alloy present on the ceramic substrate, wherein the copper or copper alloy - Grain sizes in the range of 10 µm to 300 µm, - a number distribution of grain sizes with a d 95 -value in the range of 140 µm to 250 µm and - a medium grain shape factor R a (Cu), determined as the arithmetic mean of the form factors R K the grain size of the copper or copper alloy, of at least 0.60, wherein the form factor R K a grain results from the following relationship: RK = d K ,ortho / d K ,max where d K,max the maximum grain diameter of the grain is and d K,ortho the one halfway along and perpendicular to d K,max The grain diameter varies.

Description

[0001] The present invention relates to a copper-ceramic composite and a module containing this composite, which can be used in power electronic components.

[0002] Ceramic circuit carriers are of particular interest in the field of high-performance electronics due to their high thermal conductivity, high dimensional stability and mechanical strength as well as their high insulation strength.

[0003] Various processes are available for metallizing a ceramic substrate, such as direct copper bonding (commonly referred to as DCB process), direct aluminum bonding (commonly referred to as DAB process) or active metal brazing (commonly referred to as AMB process).

[0004] The composite material obtained after metallization of the ceramic substrate is also referred to as a metal-ceramic substrate or metal-ceramic composite. If it was produced using a DCB process, for example, the term "DCB substrate" is often used.

[0005] The DCB process takes advantage of the fact that oxygen reduces the melting point of copper from 1083°C to the eutectic melting point of 1065°C. Oxidation of copper foils prior to metallization of the ceramic substrate, or the addition of oxygen during the high-temperature process (e.g., at a temperature between 1065°C and 1080°C), creates a thin eutectic melt layer. This reacts with the surface of the ceramic substrate, allowing the ceramic and metal to bond together firmly.

[0006] DCB processes are described, for example, in US 3 744 120 A or DE 23 19 854 A.

[0007] Metallization can, for example, be performed on only one side of the ceramic substrate ("single-layer bonding" SLB) or alternatively on both sides of the ceramic substrate simultaneously ("double-layer bonding" DLB). It is also possible to first metallize a first side of the substrate in a first SLB step and then metallize the opposite side of the substrate in a further SLB step.

[0008] It is also known to structure the applied metal coating for the formation of conductor tracks, for example using etching processes.

[0009] In many power electronics applications, the metal-ceramic composite is subject to high thermal cycling stresses, where significant temperature changes (e.g. in the range of -40°C to +150°C) can occur.

[0010] Due to the different thermal expansion coefficients of the ceramic substrate and the metal coating, significant mechanical stresses arise at the transition between these layers during temperature fluctuations, which can ultimately lead to at least partial detachment of the metal from the ceramic surface. It is known that specific structuring of the metal layer in its edge region can reduce the gradient of tensile and compressive stresses and thus improve thermal shock resistance. DE 40 04 844 A and DE 4318241 A1 describe metal coatings on ceramic substrates that have edge weakenings in the form of depressions or holes.

[0011] For later use in modules or components of high-performance electronics, the metal coating can be bonded with wires (often referred to as "bond wires"). This connection between the metal coating and the bond wire should be as strong as possible to minimize the risk of unwanted detachment of the bond wire.

[0012] In principle, it is desirable for the metal coating to exhibit both good adhesion to the ceramic substrate and enable good bonding of metal wires. However, since ceramic substrates and bond wires differ significantly in terms of their materials, it remains a challenge to meet both requirements simultaneously.

[0013] Other relevant properties of a copper-ceramic composite for applications in electronics are its thermal conductivity and mechanical strength (in particular the thermal conductivity and mechanical strength of the ceramic substrate) as well as the strongest possible bond between the metal coating and the ceramic surface, whereby this bond should remain sufficiently strong even under prolonged thermal cycling.

[0014] DE 10 2012 110 322 A1 defines the ceramic substrate of the metal-ceramic composite in more detail with regard to its grain structure (i.e., its structure at the microscopic level). The ceramic substrate contains a zirconium oxide-reinforced aluminum oxide, with the average grain size of the aluminum oxide being in the range of 2-8 µm and the ratio of the length of the grain boundaries of the Al2O3 grains to the total length of all grain boundaries being >0.6. According to DE 10 2012 110 322 A1, this grain structure contributes to improving thermal conductivity.

[0015] DE 42 10 900 A1 describes the production of a copper-ceramic composite, in which a copper coating is applied to a ceramic substrate by thermal spraying. A copper powder with an average particle size of less than 20 µm is used for thermal spraying.

[0016] EP 1 365 637 A2 describes the application of a copper coating to a ceramic surface or a silver-containing adhesion-promoting layer, whereby a copper powder is deposited at high velocity in a carrier gas onto the surface to be coated. If the deposition occurs on the ceramic surface, a copper powder with particles with a maximum diameter of 45 µm is used.

[0017] DE 689 27 531 T2 describes a method for producing a printed circuit board with a ceramic composite substrate, the method comprising: (a) applying one side of a copper element to a surface of a ceramic substrate, (b) heating the assembly to a temperature above 1065°C but below 1083°C, wherein the copper element has an initial surface unevenness such that the median value of the surface unevenness of the hot-bonded copper element is not greater than 3 µm and the maximum surface unevenness is not greater than 18 µm.

[0018] An object of the present invention is to provide a metal-ceramic composite with an improved property profile, in particular good thermal shock resistance and good wire bonding behavior.

[0019] The task is solved by a copper-ceramic composite, comprising - a ceramic substrate, - a coating of copper or a copper alloy present on the ceramic substrate, wherein the copper or the copper alloy - Grain sizes in the range of 10 µm to 300 µm, - a number distribution of grain sizes with a d 95 -value in the range of 140 µm to 250 µm and - an average grain shape factor R a (Cu), determined as the arithmetic mean of the form factors R K of the grains of copper or copper alloy, of at least 0.60, wherein the shape factor R K of a grain results from the following relationship: RK=dK,ortho / dK,max where d K,max is the maximum grain diameter of the grain and d K,ortho which is halfway along and perpendicular to the K,max running grain diameter.

[0020] The copper coating of a copper-ceramic composite is generally a polycrystalline material consisting of small crystallites (also called grains). At the microscopic level, polycrystalline materials can be characterized in more detail based on their grain structure (e.g., grain size distributions, grain shape, texture, etc.).

[0021] Within the scope of the present invention, it was discovered that a coating made of copper or a copper alloy with grain sizes in the range of 10 µm to 300 µm still exhibits good adhesion to a ceramic substrate even when the copper-ceramic composite is exposed to thermal cycling over extended periods. At the same time, the copper or copper alloy with these grain sizes enables efficient wire bonding.

[0022] In the context of this application, thermal shock resistance is understood to mean the resistance or resilience of the copper layer to delamination from the ceramic of a copper-ceramic substrate, with the resistance being determined as a result of at least one thermal shock of the copper layer relative to the ceramic. Improved thermal shock resistance means that the number of thermal shocks withstood is increased.

[0023] In the context of the present application, improved wire bonding is understood to mean that the force with which the bonding wire can be detached from the copper surface of a copper-ceramic composite is increased.

[0024] In the context of the present application, improved copper bond strength is understood to mean the bond strength of the copper to the copper-ceramic composite, so that the force with which the bonded copper foil can be detached from the ceramic surface of a copper-ceramic composite is increased. An exemplary measurement method is known to the person skilled in the art from DE102004012231B4 ( Fig. 2 and Fig. 3 of DE102004012231B4).

[0025] In the context of this application, improved flexural strength of the ceramic is understood to mean that the force leading to fracture is increased during three-point bending. For example, the person skilled in the art will be familiar with DIN EN 843-1 (2008) for determining the flexural strength of ceramics. Preferably, the sample geometry differs from DIN EN 843-1 (2008) in that the samples have dimensions of 20 × 40 × 0.38 mm. 3 or 20 × 40 × 0.63 mm 3 have.

[0026] The grain sizes of the copper or copper alloy are preferably in the range of 15 µm to 250 µm, more preferably in the range of 20 µm to 210 µm.

[0027] In the present case, these values ​​are not to be regarded as strict lower and upper limits for the grain size distribution, but can vary by + / - 10%. In a preferred embodiment, however, they are the lower limit, which is not undercut, and the upper limit, which is not exceeded, of the grain size distribution of the copper or copper alloy. In this preferred embodiment, therefore, the copper or copper alloy does not have any grains that lie outside the above-mentioned ranges. It is therefore preferred that the min (Cu) ≥ 10 µm and d max (Cu) ≤ 300 µm, preferred d min (Cu) ≥ 15 µm and d max (Cu) ≤ 250 µm, more preferably d min (Cu) ≥ 20 µm and d max (Cu) ≤ 210 µm, where d min (Cu) and d max(Cu) are the minimum and maximum grain sizes of copper.

[0028] By using a copper starting foil with a suitable grain size distribution, the desired grain sizes in the copper-ceramic composite can be adjusted. Such copper foils are commercially available or can be obtained using standard processes. Fine adjustment of the grain sizes can be achieved, if necessary, by thermal treatment of the starting foil.

[0029] As is generally known to those skilled in the art, the grain size distribution can be determined by reference to the number of grains (i.e., number distribution) or, alternatively, by reference to the mass (i.e., mass distribution) or the volume of the grains. Within the scope of the present invention, the grain size distribution is determined based on the number of grains.

[0030] In a preferred embodiment, the copper or copper alloy has a number distribution of grain sizes in which at most 5% of the grains have a grain size of less than 15 µm, preferably less than 20 µm, more preferably less than 25 µm; and / or in which at least 95% of the grains have a grain size of less than 250 µm, preferably less than 230 µm, more preferably less than 200 µm.

[0031] As is well known, characteristic values ​​of a grain size distribution include its d 50 -value, d5-value and d 95 -value. For the d 50 -value, which is often also referred to as the median value, the following applies: 50% of the grains have a diameter that is smaller than the d 50 -Value.

[0032] Similarly, for the d5 value, 5% of the grains have a diameter that is smaller than this d5 value, and for the d 95-value means that 95% of the grains have a diameter that is smaller than this d 95 -Value.

[0033] The arithmetic mean d arith A grain size distribution results from the sum of the grain sizes of the individual grains divided by the number of grains.

[0034] Within the scope of the present invention, it was surprisingly found that a further optimization of the thermal shock resistance and the bonding behavior of the copper or copper alloy can be achieved if the d5 and d 95 -values ​​of the grain size distribution meet certain requirements.

[0035] The number distribution of the grain sizes of the copper or copper alloy shows a d 95The d5 value is in the range of 140 µm to 250 µm, more preferably in the range of 140 µm to 230 µm, and even more preferably in the range of 150 µm to 200 µm. The d5 value of the grain size number distribution of the copper or copper alloy is preferably ≥ 15 µm; more preferably, the d5 value is in the range of 15 µm to 80 µm, more preferably in the range of 20 µm to 75 µm, and even more preferably in the range of 25 µm to 70 µm.

[0036] A preferred d 50 -value of the grain size distribution of the copper or copper alloy is, for example, in the range of 55 µm to 115 µm.

[0037] Within the scope of the invention, it was further discovered that product properties could be improved by selecting additional product parameters.

[0038] In the present case, it may be preferred that the d5, d 95 - and d 50-Values ​​of the grain size distribution of the copper or copper alloy are chosen so that they satisfy the following condition: 4.0≥(d95−d5) / d50≥0.5

[0039] The symmetry of a grain size distribution can be determined by the ratio of median value d 50 to the arithmetic mean d arith this distribution (ie by the quotient d 50 / d arith ; hereinafter also referred to as symmetry value S). The closer the symmetry value is to 1.0, the more symmetrical the grain size distribution. In a preferred embodiment, the copper or copper alloy has a number distribution of grain sizes with a median value d 50 and an arithmetic mean d arith where the ratio of d 50 to the arith (ie d 50 / d arith) is in the range from 0.75 to 1.10, more preferably in the range from 0.78 to 1.05, even more preferably in the range from 0.80 to 1.00. This allows further optimization of the thermal shock resistance and the wire bonding properties. Suitable methods with which the symmetry of the grain size distribution in the copper can be adjusted, for example already in the copper starting foil, are known to the person skilled in the art. For example, the symmetry of the grain size distribution in a copper foil can be influenced by a suitable processing temperature or a rolling process. Copper starting foils with which the above-mentioned symmetry values ​​can be achieved in the final copper-ceramic composite are commercially available or can be obtained using standard processes.

[0040] The width of the grain size distribution can be determined by the ratio of d5 value to d 95-value. In a preferred embodiment, the copper or copper alloy has a number distribution of grain sizes with a d5 value and a d 95 -value, where the ratio of d5 to d 95 in the range of 0.1 to 0.4, more preferably in the range of 0.11 to 0.35, even more preferably in the range of 0.12 to 0.30. This allows for further optimization of thermal shock resistance and wire bonding properties.

[0041] Further optimization of thermal shock resistance can be achieved if the grains of the copper coating, in their two-dimensional projection in a plane parallel to the surface of the ceramic substrate, have a shape that is as circular or round as possible. The shape of an individual grain can be determined by its shape factor R K which is the ratio of the maximum grain diameter d K,max to the perpendicular to d K,maxrunning grain diameter d K,ortho , determined at half the length of d K,max , is (ie R K = d K,ortho / d K,max ). From the arithmetic mean of the form factors R K of the grains, the mean grain shape factor R is obtained a (Cu) of the copper or copper alloy. For example, if a copper material contains a high proportion of elongated grains, the average grain shape factor of this copper will be relatively low. On the other hand, the higher the proportion of round, circular grains, the closer the average grain shape factor approaches 1.0.

[0042] The mean grain shape factor R a(Cu) of the copper or copper alloy is ≥0.60, more preferably ≥0.80. Suitable methods for adjusting the shape of the grains in the copper, for example, in the copper starting foil, are known to those skilled in the art. For example, the grain shape in a copper foil can be influenced by a suitable processing temperature or a rolling process. Copper starting foils with which the above-specified average grain shape factor R can be achieved in the final copper-ceramic composite a (Cu) are commercially available or can be obtained via standard procedures.

[0043] A suitable thickness of the coating of copper or a copper alloy in a copper-ceramic composite is known to those skilled in the art. As explained below, at some points on the coating, particularly in edge regions, a portion of the copper or copper alloy can be removed again, for example, to form edge weakenings. Therefore, within the scope of the present invention, it is possible for the thickness of the metal coating to vary. Typically, the coating of copper or a copper alloy has a thickness in the range of 0.2-1.2 mm over at least 70% of its surface. For example, it is possible for the thickness to be approximately 300 µm.

[0044] Preferably, the thickness of the coating is made of copper or a copper alloy (D Cu ) and the median value d 50 the grain size distribution of the copper or copper alloy is selected so that the ratio D Cu to the 50in the range of 0.05 to 0.40. For this purpose, the thickness D Cu of the copper or copper alloy at a point of the coating and divided by the median value d 50 the grain size distribution of the copper or copper alloy. Preferably, the ratio D Cu / d 50 over at least 70%, more preferably at least 90% of the area of ​​the coating made of copper or a copper alloy in the range of 0.05 to 0.40.

[0045] Preferably, the copper of the coating has a purity of ≥ 99.50%, more preferably ≥ 99.90%, even more preferably ≥ 99.95% or even ≥ 99.99%.

[0046] Preferably, the copper or copper alloy coating is applied to the ceramic substrate using a DCB process. As already explained above, a typical DCB process may include the following process steps: - oxidizing a copper foil so that a copper oxide layer forms on its surface; - Placing the copper foil with the copper oxide layer on the ceramic substrate; - Heating the composite to a temperature < 1083°C (e.g. a temperature in the range of 1065-1080°C), - Cool to room temperature.

[0047] As a result of the DCB process, spinel crystallites (e.g. copper-aluminum spinels) may be present between the copper or copper alloy coating and the ceramic substrate.

[0048] The coating of copper or a copper alloy can, for example, be applied only to one side of the ceramic substrate. Alternatively, it is possible for both sides (i.e., top and bottom) of the ceramic substrate to be provided with the coating of copper or a copper alloy. An exemplary copper-ceramic composite, in which a ceramic substrate 1 has a coating 2 of copper or a copper alloy on both its bottom and top sides, is described in Fig. 1. An exemplary copper-ceramic composite, in which the ceramic substrate 1 has several regions, each provided with a coating 2 of copper or copper alloy, is shown in Fig. 2. As will be explained below, the individual metallized areas can be separated by predetermined breaking lines (not shown in Fig.2) be separated from each other so that these areas can be separated by breaking along these predetermined breaking lines.

[0049] To form electrical contact surfaces, the coating made of copper or a copper alloy can be at least partially structured. The structuring of the metal coating can be carried out in a known manner, in particular by an etching process (for example, using an etching mask).

[0050] In the etching process, the copper or copper alloy can be completely removed in partial areas, so that the surface of the ceramic substrate is exposed in these partial areas. Furthermore, it is also possible for the coating of copper or a copper alloy to have one or more depressions (preferably round depressions), which are obtained, for example, in the etching process by only partially removing the copper or copper alloy in the area of ​​the depression to be made, and the surface of the ceramic substrate in this area is therefore still covered with copper or copper alloy. Alternatively or additionally, it is possible to etch the depressions through the copper or copper alloy down to the ceramic surface.With regard to the possible arrangement of such depressions, preferably in the edge region of the coatings made of copper or copper alloy, reference can be made, for example, to DE 40 04 844 C1 and DE 43 18 241 A1.

[0051] Suitable materials for the ceramic substrate are known to those skilled in the art. The ceramic substrate preferably contains an oxide, a nitride, a carbide, or a mixture or composite of at least two of these materials. The ceramic substrate can consist of only one layer or, alternatively, of multiple layers.

[0052] Suitable oxides include aluminum oxide (Al2O3) or BeO. If the oxide is aluminum oxide, it can optionally be reinforced with zirconium oxide (ZrO2). Such a ZrO2-reinforced Al2O3 typically contains zirconium oxide in a proportion of 0.5–30 wt% based on its total mass. The zirconium oxide, in turn, can optionally be doped with one or more doping oxides, in particular yttrium oxide, calcium oxide, cerium oxide, or magnesium oxide, typically in a proportion of up to 0.01 wt% or even up to 5 wt%, based on the total mass of zirconium oxide and aluminum oxide.

[0053] If the ceramic substrate contains a nitride, this can be, for example, an aluminum nitride, a silicon nitride, or a titanium nitride. For example, the ceramic substrate can be a multilayer ceramic substrate containing at least one layer of aluminum nitride and at least one layer of Al2O3, with the coating of copper or a copper alloy applied to the Al2O3 layer.

[0054] If the ceramic substrate contains a carbide, this can be, for example, silicon carbide.

[0055] In a preferred embodiment, the ceramic substrate contains aluminum oxide (Al2O3). As already mentioned above, this can be Al2O3 reinforced with ZrO2.

[0056] The ceramic substrate preferably contains at least 65 wt% Al2O3. If no ZrO2 is present to reinforce the Al2O3, the ceramic substrate can, for example, consist of at least 95 wt%, preferably 96 wt%, of Al2O3.

[0057] If a ZrO2-reinforced aluminum oxide is used (where the ZrO2 is optionally doped, as mentioned above), the ceramic substrate can, for example, consist of at least 96 wt%, preferably at least 98 wt%, of this ZrO2-reinforced Al2O3.

[0058] The grain sizes of the aluminum oxide are preferably in the range from 0.01 µm to 25 µm, more preferably in the range from 0.3 µm to 23 µm, and even more preferably in the range from 0.5 µm to 20 µm. With grain sizes in this range, the ceramic substrate of the copper-ceramic composite has both high mechanical strength and high thermal conductivity. These values ​​are not to be regarded as strict lower and upper limits for the grain size distribution, but can vary by + / - 10%. In a preferred embodiment, however, these are the lower limit, which is not undershot, and upper limits, which are not exceeded, of the grain size distribution of the aluminum oxide. In this preferred embodiment, therefore, the aluminum oxide does not have any grains that lie outside the above-mentioned ranges. It is therefore preferred that the min (Al2O3) ≥ 0.01 µm and d max (Al2O3) ≤ 25 µm, preferred d min (Al2O3) ≥ 0.3 µm and d max(Al2O3) ≤ 23 µm, more preferably d min (Al2O3) ≥ 0.5 µm and d max (Al2O3) ≤ 20 µm, where d min (Al2O3) and d max (Al2O3) are the minimum and maximum grain sizes of the aluminum oxide.

[0059] By using a ceramic starting material with a suitable Al2O3 grain size distribution, the desired Al2O3 grain sizes in the copper-ceramic composite can be adjusted. Such ceramic materials are commercially available or can be obtained using standard processes. Fine adjustment of the grain sizes can be achieved, if necessary, by thermal treatment of the ceramic starting material.

[0060] As already mentioned above, in the present invention, the distribution of grain sizes is determined on the basis of the number of grains (ie, grain size number distribution).

[0061] In a preferred embodiment, the aluminum oxide of the ceramic substrate has a number distribution of grain sizes in which at most 5% of the grains have a grain size of less than 0.1 µm, more preferably less than 0.3 µm, even more preferably less than 0.5 µm; and / or in which at least 95% of the grains have a grain size of less than 15 µm, more preferably less than 10 µm, even more preferably less than 7 µm.

[0062] An improvement or optimization of the mechanical strength and thermal conductivity of the ceramic substrate in the metal-ceramic composite can be achieved if the d5 and d 95 -Values ​​of the grain size distribution of Al2O3 meet certain requirements.

[0063] Preferably, the number distribution of the grain sizes of the aluminum oxide has a d 95-value of ≤ 15.0 µm, more preferably in the range from 4.0 µm to 15.0 µm, even more preferably in the range from 4.5 µm to 10.0 µm, even more preferably in the range from 5.0 µm to 8.0 µm. The d5 value of the grain size number distribution of the aluminum oxide is preferably ≥ 0.1 µm; more preferably the d5 value is in the range from 0.1 µm to 2.5 µm, even more preferably in the range from 0.3 µm to 2.5 µm, even more preferably in the range from 0.5 µm to 2.0 µm. This allows further optimization of the mechanical strength and thermal conductivity of the ceramic substrate in the metal-ceramic composite.

[0064] A preferred d 50 -value of the grain size distribution of aluminum oxide, for example, is in the range of 1.0 µm to 3.0 µm.

[0065] In the present case, it may be preferred that the d5, d95 and d50 values ​​of the particle size distribution of the aluminum oxide are selected so that they satisfy the following condition: 9.5≥(d95−d5) / d50≥0.7

[0066] In a preferred embodiment, the aluminum oxide has a number distribution of grain sizes with a median value d 50 and an arithmetic mean d arith where the ratio of d 50 to the arith (ie d 50 / d arith ; hereinafter also referred to as the symmetry value S(Al2O3) of the grain size distribution of the aluminum oxide)) is in the range of 0.75 to 1.10, more preferably in the range of 0.78 to 1.05, even more preferably in the range of 0.80 to 1.00. This allows for further optimization of the mechanical strength and thermal conductivity of the ceramic substrate in the metal-ceramic composite.

[0067] Suitable methods for adjusting the symmetry of the grain size distribution in aluminum oxide, for example, during the production of the starting substrate, are known to those skilled in the art. For example, the symmetry of the grain size distribution can be influenced by the sintering time and sintering temperature during the production of the starting substrate. Al2O3 substrates that can achieve the above-mentioned symmetry values ​​in the final copper-ceramic composite are commercially available or can be obtained using standard processes.

[0068] The width of the grain size distribution can be determined by the ratio of d5 value to d 95 -value. In a preferred embodiment, the aluminum oxide has a number distribution of grain sizes with a d5 value and a d 95 -value, where the ratio of d5 to d 95in the range of 0.1 to 0.4, more preferably in the range of 0.11 to 0.35, even more preferably in the range of 0.12 to 0.30. This allows for further optimization of the mechanical strength and thermal conductivity of the ceramic substrate in the metal-ceramic composite.

[0069] The preferred grain shape factor is R a (Al2O3) of the aluminum oxide ≥0.40, more preferably ≥0.60, even more preferably ≥0.80. As explained above, the shape of an individual grain can be determined by its shape factor R K which is the ratio of the maximum grain diameter d K,max to the perpendicular to d K,max running grain diameter d K,ortho , determined at half the length of d K,max , is (ie R K = d K,ortho / d K,max ). From the arithmetic mean of the form factors R K of the grains, the mean grain shape factor R is obtained a(Al2O3) of the aluminum oxide. Suitable methods for adjusting the shape of the aluminum oxide grains, for example, during the production of the starting substrate, are known to those skilled in the art. For example, the shape of the Al2O3 grains can be influenced by the sintering time and sintering temperature during the production of the starting substrate. Al2O3 substrates with which the above-mentioned shape factor R can be achieved in the final copper-ceramic composite a (Al2O3) are commercially available or can be obtained via standard processes.

[0070] A suitable thickness of the ceramic substrate in a copper-ceramic composite is known to those skilled in the art. Typically, the ceramic substrate has a thickness in the range of 0.2-1.2 mm over at least 70% of its surface area, more preferably at least 90% of its surface area. A thickness of the ceramic substrate is, for example, approximately 0.38 mm or approximately 0.63 mm.

[0071] Preferred are the thickness of the ceramic substrate (D cer ) and the median value d 50 the grain size distribution of the aluminum oxide in the ceramic substrate is selected so that the ratio D cer to the 50 (ie D cer / d 50 ) is in the range of 0.001 to 0.01, more preferably in the range of 0.002 to 0.009, even more preferably in the range of 0.004 to 0.008. For this purpose, the thickness D cer of the ceramic substrate at one point and divided by the median value d 50 the particle size distribution of the aluminum oxide. Preferably, the ratio D cer / d 50 over at least 70%, more preferably at least 90% of the area of ​​the ceramic substrate in the range of 0.05 to 0.40.

[0072] The ceramic substrate can, for example, have a thermal conductivity of ≥ 20 W / mK, and / or a flexural strength of ≥ 400 MPa.

[0073] The ceramic substrate can be in the form of a single substrate. Alternatively, it is also possible for the ceramic substrate to have one or more (preferably straight) predetermined breaking lines that divide the ceramic substrate into two or more regions, with the coating of copper or a copper alloy applied to at least one of these regions. Regarding the structure of such a multiple substrate with predetermined breaking lines, reference can be made, for example, to DE 43 19 944 A1 and DE 199 27 046 A1.

[0074] Suitable dimensions (length × width) of the ceramic substrate (either as a single substrate or as a multiple substrate) in a metal-ceramic composite are known to those skilled in the art. For example, the ceramic substrate can have a dimension (length × width) of (180-200 mm) × (130-150 mm) or (180-200 mm) × (270-290 mm). Smaller dimensions, for example, (8-12 mm) × (8-12 mm), are also possible.

[0075] A further improvement in the properties of the copper-ceramic composite can be achieved if the grain properties of the copper or copper alloy and the grain properties of the aluminum oxide in the ceramic substrate are matched.

[0076] In a preferred embodiment, the ratio of d 50 (Al2O3) to d 50 (Cu) in the range of 0.008 to 0.055, more preferably in the range of 0.010 to 0.045. This allows for further optimization of the metal-ceramic adhesion and thermal shock resistance in the metal-ceramic composite.

[0077] In a preferred embodiment, the copper or copper alloy has a number distribution of grain sizes with a median value d 50 , an arithmetic mean d arith and a symmetry value S(Cu) = d 50 / d arith The aluminum oxide has a number distribution of grain sizes with a median value d50 , an arithmetic mean d arith and a symmetry value S(Al2O3)= d 50 / d arith where S(Al2O3) and S(Cu) satisfy the following condition: 0.7≤S(Al2O3) / S(Cu)≤1.4.

[0078] Preferably, S(Al2O3) and S(Cu) satisfy the following condition: 0.74≤S(Al2O3) / S(Cu)≤1.35; even more preferably the following condition 0.80≤S(Al2O3) / S(Cu)≤1.25.

[0079] This improves the thermal shock resistance of the copper-ceramic composite.

[0080] In a preferred embodiment, the average grain shape factor of the aluminum oxide R a (Al2O3) and the average grain shape factor of the copper or copper alloy R a (Cu) of the following condition: 0.5≤R a (Al2O3) / R a (Cu) ≤2.0.

[0081] Even more preferred is 0.75≤Ra(Al2O3) / Ra(Cu)≤1.5 and even more preferred is 0.80≤Ra(Al2O3) / Ra(Cu)≤1.20.

[0082] This allows for further optimization of the thermal shock resistance of the copper-ceramic composite.

[0083] In a preferred embodiment, the aluminum oxide has grain sizes in the range of d min (Al2O3) to d max (Al2O3), the copper or copper alloy has grain sizes in the range of d min (Cu) to d max (Cu), and the ratios of d min (Al2O3) to d max (Cu) and from d max (Al2O3) to d min (Cu) satisfy the following conditions (i) and (ii): dmin(Al2O3) / dmax(Cu)≥1×10−5 and 2.5≥dmax(Al2O3) / dmin(Cu).

[0084] Even more preferably, the ratios of d min (Al2O3) to d max (Cu) and from d max (Al2O3) to d min (Cu) the following conditions (i) and (ii): dmin(Al2O3) / dmax(Cu)≥0.001 and 1.5≥dmax(Al2O3) / dmin(Cu) and most preferably the following conditions (i) and (ii): dmin(Al2O3) / dmax(Cu)≥0.002 and 1.0≥dmax(Al2O3) / dmin(Cu).

[0085] In a particularly preferred embodiment, 0.005≥dmin(Al2O3) / dmax(Cu)≥0.002 and 1.0≥dmax(Al2O3) / dmin(Cu)≥0.05.

[0086] This allows a strong bond to be created between the metal coating and the ceramic substrate, which can withstand even frequent thermal cycling. As mentioned above, it is preferred that the min (Cu) ≥ 10 µm and d max (Cu) ≤ 300 µm, preferred d min (Cu) ≥ 15 µm and d max (Cu) ≤ 250 µm, more preferably d min (Cu) ≥ 20 µm and d max (Cu) ≤ 210 µm, where d min (Cu) and d max(Cu) are the minimum and maximum grain sizes of copper.

[0087] Furthermore, the present invention relates to a module comprising at least one copper-ceramic composite as described above and one or more bonding wires. Typically, the bonding wire or wires are bonded to the copper or copper alloy coating. Suitable bonding methods for connecting wires to a metal coating are known to those skilled in the art. The module may also contain one or more electronic components, such as one or more chips.

[0088] The grain structures of the copper or copper alloy and the aluminum oxide of the ceramic substrate are determined as follows in the context of the present application: Grain size distribution of the copper or copper alloy

[0089] A light micrograph is taken of the surface of the copper or copper alloy coating (parallel to the coated substrate surface). Special sample preparation in the form of a micrograph is not required.

[0090] The grain sizes are determined using a line-intercept method. Line-intercept methods are known to those skilled in the art and are described, for example, in ASTM 112-13.

[0091] The magnification is selected so that at least 50 grains are cut through the line grid.

[0092] In this example, two parallel lines in the x-direction and two parallel lines in the y-direction were placed in the light microscopic image. The lines divide the image into three equally wide stripes. For illustration, this is shown schematically in Fig. 3 shown.

[0093] If a grain is intersected by one of these lines over a length L, this length L is taken as the grain size. For each grain intersected by one of these lines, a grain size is obtained. At the intersection point of two lines, two values ​​are obtained for a grain, both of which are used to determine the grain size distribution.

[0094] The grain sizes of the cut grains result in a grain size distribution, from which the d5, d 50 - and d 95 -values ​​and the arithmetic mean d arith can be determined. As already explained above and is generally known to the person skilled in the art, the following applies to the d 50 -value, which is often also referred to as the median value, is as follows: 50% of the grains have a diameter that is smaller than the d 50 -value. Analogously, for the d5 value, 5% of the grains have a diameter that is smaller than this d5 value, and for the d 95-value means that 95% of the grains have a diameter that is smaller than this d 95 -Value. The arithmetic mean of the grain size distribution is the sum of the grain sizes of the individual grains divided by the number of cut grains. Grain size distribution of the aluminum oxide of the ceramic substrate

[0095] A scanning electron micrograph (SEM) is taken of the surface of the ceramic substrate. Special sample preparation in the form of a micrograph is not required. The SEM image is taken at a location on the ceramic substrate that was previously coated with copper and exposed by etching.

[0096] The grain sizes are determined using a line-intercept method. Line-intercept methods are known to those skilled in the art and are described, for example, in ASTM 112-13.

[0097] The magnification is selected so that at least 50 Al2O3 grains are intersected by the line grid. If the ceramic substrate contains grains of a different chemical composition, such as ZrO2 grains, these can be easily distinguished from the Al2O3 grains in the SEM image using secondary electron contrast and are therefore not included in subsequent calculations.

[0098] With regard to further evaluation using the line cutting method, reference can be made to the explanations given above for copper coating.

[0099] Both the grain sizes of the copper or copper alloy and the grain sizes of the Al2O3 are determined in a plane that is parallel to the coated substrate surface or coplanar with it. Single grain form factor, medium grain form factor copper, copper alloy

[0100] The light microscopic image is used, which was also used to determine the grain size distribution.

[0101] The following procedure is used to determine the shape factor of an individual grain: Its longest dimension is d K,max Then, halfway along the length of the K,max which is perpendicular to d K,max running diameter d K,ortho of the grain. The shape factor of the individual grain R K results from the ratio of d K,ortho to the K,max , i.e. R K = d K,ortho / d K,max . This is shown schematically in Fig.4 is illustrated using a grain with an elliptical grain structure. The more the shape of a grain in its two-dimensional projection approaches that of a circle, the more the grain's shape factor approaches the value 1.0. The shape factor is therefore also a measure of the circularity / roundness of the grains.

[0102] The shape factor is determined for at least 50 grains from the light micrograph. Typically, the grains that were also intersected by the lines in the line-intersecting method are evaluated.

[0103] The mean grain shape factor of the copper or copper alloy is then calculated from the arithmetic mean of the shape factors of the individual grains (i.e. sum of the individual shape factors divided by the number of grains examined). aluminum oxide

[0104] The SEM image is used, which was also used to determine the grain size distribution.

[0105] Regarding the determination of the shape factor of individual grains as well as the average grain shape factor of Al2O3, reference can be made to the explanations given above for copper.

[0106] Both the grain shape factors of the copper or the copper alloy as well as the grain shape factors of the Al2O3 are determined in a plane that is parallel to the coated substrate surface or coplanar with it.

[0107] The following describes a bonding process which is preferably used to produce the copper-ceramic substrates according to the invention: A typical process which is preferably used in the context of the present invention for applying the copper coating to the ceramic substrate is known, for example, from the documents US 3 744 120 A, US 3 994 430 A, EP 0 085 914 A2 or DE 23 19 854 A, the corresponding disclosure of which is incorporated into the present invention by reference.

[0108] The manufacturing processes disclosed therein, for example, in the context of the direct copper bonding process (DCB process), have in common that a copper foil is first oxidized to form a substantially uniform copper oxide layer. The resulting copper foil is then positioned on a ceramic substrate, and the composite of ceramic substrate and copper foil is heated to a process or bonding temperature between approximately 1025 and 1083 °C, resulting in the formation of a metallized ceramic substrate. After bonding, the copper foil thus represents a coating. Finally, the resulting metallized ceramic substrate is cooled.

[0109] The ceramic substrate and copper foil are bonded together in a furnace, generally known as a bonding furnace. Such bonding furnaces, often also called tunnel furnaces, comprise, among other things, an elongated, tunnel-like furnace chamber (also called a muffle) and a transport device with a transport element, for example in the form of a flexible, heat-resistant conveyor belt, for transporting the material to be treated through the furnace chamber, which is heated by a heating device. The ceramic substrates, together with the copper foil, are positioned on a carrier on the conveyor belt and then, driven by the conveyor belt, pass through a heating area in the bonding furnace, where the required bonding temperature is reached. At the end of the bonding process, the resulting composite of ceramic substrate and copper foil is cooled again.

[0110] This process can be used for the production of single-sided metallized ceramic substrates as well as for the production of double-sided metallized substrates. The production of double-sided metallized substrates is generally carried out using a two-stage bonding process, i.e., a two-stage single-layer bonding process (SLB process). In this case, a two-stage bonding process is preferred.

[0111] In this two-step bonding process for producing double-sided metallized ceramic substrates, the ceramic is bonded to the copper foils on the opposite sides of the ceramic substrate in two furnace passes.

[0112] For this purpose, a ceramic substrate is first placed on a carrier and then covered with a copper foil on the upper side, i.e., the side facing away from the carrier. This side of the ceramic substrate is bonded to the metal layer by applying heat, and the resulting assembly is then cooled.

[0113] The substrate is then turned over and in a second bonding step the other side of the substrate is provided with a metal layer, i.e. the copper foil, in the same way.

[0114] It is possible to produce individual part cards or large cards that have several break-out individual cards. Examples

[0115] The following examples show how the grain size distribution in the copper coating influences the thermal shock resistance and the wire bonding behavior of a copper-ceramic composite.

[0116] Three copper-ceramic samples were produced using a DCB process, which differed in their grain size distributions: Copper-ceramic composite 1, hereinafter referred to as “KKV 1” (according to the invention) Copper-ceramic composite 2, hereinafter referred to as “KKV 2” (comparison sample) Copper-ceramic composite 3, hereinafter referred to as “KKV 3” (comparison sample)

[0117] In each of these three copper-ceramic composites, both the top and bottom surfaces of the ceramic substrate were coated with a copper coating. The copper coating was first bonded to one side of the ceramic substrate using the SLB process. Subsequently, the opposite side of the ceramic substrate was coated with another copper coating using the SLB process, creating a copper-ceramic substrate with a copper foil bonded to both sides of the ceramic. One of the two copper coatings was subsequently patterned in each sample using an etching process (the same patterning was used for all samples).

[0118] In each of these 3 copper-ceramic composites, the thickness of the copper coating was 0.3 mm and the length × width of the copper coating was: 181 × 132 mm 2 . In the example shown, it is pure copper.

[0119] Furthermore, all 3 samples contained an Al2O3 ceramic substrate (Al2O3 content: 96 wt%), thickness of the ceramic substrate: 0.38 mm; length × width of the ceramic substrate: 190 × 140 mm 2

[0120] The selection of the copper coating was chosen so that the inventive effect, namely a good thermal shock resistance after the two-stage SLB bonding (copper-ceramic substrate according to the invention) and a good bonding of wires on the metal coating after the two-stage SLB bonding (copper-ceramic substrate according to the invention), was realized.

[0121] The copper coating of KKV 1 had grain sizes ranging from 32 µm (i.e. lower limit of the grain size distribution) to 192 µm (i.e. upper limit of the grain size distribution). The d5 value of the grain size distribution was 49 µm and the d 95 -value was 177 µm. The d 50 -value of the grain size distribution was 98 µm.

[0122] Fig.Figure 5 shows a light microscopic image of the surface of the copper coating of KKV 1, by means of which the copper grain sizes of the KKV-1 embodiment were determined.

[0123] The copper coating of KKV 2 had grain sizes ranging from 4 µm (i.e. lower limit of the grain size distribution) to 175 µm (i.e. upper limit of the grain size distribution). The d5 value of the grain size distribution was 13 µm and the d 95 -value was 154 µm. The d 50 -value of the grain size distribution was 42 µm.

[0124] The copper coating of KKV 3 had grain sizes ranging from 43 µm (i.e. lower limit of the grain size distribution) to 470 µm (i.e. upper limit of the grain size distribution). The d5 value of the grain size distribution was 72 µm and the d 95 -value was 293 µm. The d 50 -value of the grain size distribution was 156 µm.

[0125] For each of these 3 samples, both the thermal shock resistance and the wire bonding properties were determined.

[0126] The thermal shock resistance was determined as follows: To determine the thermal shock resistance of the copper-ceramic substrate, a single substrate section is preferably cut from a large card. The single substrate section was subjected to a thermal shock cycle in a device known to those skilled in the art, which consists of: • Storage at 150 °C (preferably in the first chamber of a temperature change cabinet) for 15 minutes • Storage at - 40 °C (minus 40 °C) (preferably in a second chamber of the temperature change cabinet) for 15 minutes • , with a transfer time of 15 seconds between the movement from one chamber to the other.

[0127] During 5 cycles (aging from 150 °C to -40 °C and back corresponds to one cycle), the surface of the connection at the interface between copper and ceramic was examined for delamination using an ultrasonic microscope.

[0128] In the case of the sample KKV 1 according to the invention, a significantly reduced delamination was observed compared to the samples KKV 2 and KKV 3.

[0129] The wire bonding properties were determined as follows: To determine the wire bonding properties of the copper-ceramic substrate, a single substrate is preferably cut from a large card. A wire (in the exemplary embodiment, for example, a copper wire according to a method known to those skilled in the art) was bonded to the copper surface of the bonded copper-ceramic substrate using a device known to those skilled in the art. Within the scope of the invention, it was found that, with the inventive selection of the product parameters and the subsequent DCB bonding process, sample KKV 1 exhibited significantly improved wire bonding properties compared to samples KKV 2 and KKV 3.

[0130] The results are summarized in the following table: Table: Thermal shock resistance and wire bonding properties of copper-ceramic composites Resistance to temperature changes Wire bonding Sample KKV 1 (according to the invention) Grain size range: 32-192 µm + + d5 value: 49 µm d95 value: 177 µm d50 value: 98 µm Sample KKV 2 (comparison) Grain size range: 4-175 µm - + d5 value: 13 µm d95 value: 154 µm d50 value: 42 µm Sample KKV 3 (comparison) Grain size range: 43-470 µm + - d5 value: 72 µm d95 value: 293 µm d50 value: 156 µm

[0131] As the examples demonstrate, it is only possible with the particle size distribution according to the invention, in particular by adhering to the d5 and d 95 -values, a simultaneous improvement in thermal shock resistance and wire bonding.

Claims

[1] Copper-ceramic composite, comprising - a ceramic substrate, - a coating of copper or a copper alloy present on the ceramic substrate, wherein the copper or the copper alloy - Grain sizes in the range of 10 µm to 300 µm, - a number distribution of grain sizes with a d 95 -value in the range of 140 µm to 250 µm and - an average grain shape factor R a (Cu), determined as the arithmetic mean of the form factors R K of the grains of copper or copper alloy, of at least 0.60, wherein the shape factor R K of a grain results from the following relationship: RK=dK,ortho / dK,max where d K,max is the maximum grain diameter of the grain and d K,ortho which is halfway along and perpendicular to the K,max running grain diameter. [2] Copper-ceramic composite according to claim 1, wherein the copper or the copper alloy has a number distribution of grain sizes in which at most 5% of the grains have a grain size of less than 15 µm; and / or in which at least 95% of the grains have a grain size of less than 250 µm. [3] Copper-ceramic composite according to claim 1 or 2, wherein the copper or the copper alloy has a number distribution of grain sizes with a median value d 50 in the range of 55 µm to 115 µm. [4] Copper-ceramic composite according to one of the preceding claims, wherein the copper or the copper alloy has a number distribution of grain sizes with a d 95 -value in the range of 140 µm to 230 µm. [5] Copper-ceramic composite according to one of the preceding claims, wherein the copper or the copper alloy has a number distribution of grain sizes with a d5 value of ≥ 15 µm. [6] Copper-ceramic composite according to one of the preceding claims, wherein the grain sizes of the copper or copper alloy are in the range of 15 µm to 250 µm. [7] Copper-ceramic composite according to one of the preceding claims, wherein the coating of copper or a copper alloy has a thickness D Cu and the ratio D Cu to the median value d 50 the grain size distribution of copper is in the range of 0.05 to 0.

40. [8] Copper-ceramic composite according to one of the preceding claims, wherein the coating of copper or a copper alloy is applied to the ceramic substrate by a DCB process. [9] Copper-ceramic composite according to one of the preceding claims, wherein the coating of copper or a copper alloy has at least partially a structuring for forming electrical contact surfaces. [10] Copper-ceramic composite according to one of the preceding claims, wherein the ceramic substrate contains an oxide, a nitride, a carbide, or a mixture or composite of at least two of these materials. [11] Copper-ceramic composite according to one of the preceding claims, wherein the ceramic substrate consists of at least 65 wt% Al2O3. [12] Copper-ceramic composite according to one of the preceding claims, wherein the coating of copper or a copper alloy has a thickness in the range of 0.2-1.2 mm over at least 70% of its area; and / or the ceramic substrate has a thickness in the range of 0.2-1.2 mm over at least 70% of its area. [13] Module comprising at least one copper-ceramic composite according to one of claims 1-12 and one or more bonding wires.

Images