Carrier substrate for electrical components and method for producing such a carrier substrate

Description

[0001] The present invention relates to a support substrate for electrical components and a method for producing such a support substrate.

[0002] Heat sinks for cooling electrical and electronic components, especially semiconductors such as laser diodes, are well-known in the art. During operation, these components generate heat, which is dissipated by the heat sink to ensure their continued functionality. This is particularly important for laser diodes, where even temperature differences of just a few degrees Celsius can significantly impair performance and / or lifespan.

[0003] To cool the components, the heat sinks, which are usually attached to the components, typically have a cooling fluid channel system. During operation, a cooling fluid is circulated through this channel to absorb and dissipate heat emanating from the electrical or electronic component. Preferably, a fin structure is used in which several rib-like elements project into the cooling fluid channel system to provide the largest possible contact area with the cooling fluid. This improves the heat transfer from the walls that define or project into the cooling fluid channel system to the cooling fluid. For example, these heat sinks are used to cool laser diodes.

[0004] Furthermore, it is common to use heat sinks to cool printed circuit boards designed as metal-ceramic substrates.

[0005] To insulate the electrical components on the component side, insulating elements are typically embedded in the substrates into which the heat sinks are integrated. Ceramic elements have proven particularly advantageous due to their high insulating strength. However, the choice of material for the ceramic elements used for electrical insulation is limited by the manufacturing process of the substrates, especially by the bonding of the heat sink. Therefore, for example, the positive thermal properties of a Si3N4 substrate cannot be utilized for such substrates.

[0006] German patent DE 10 2020 111 700 A1 relates to a metal-ceramic substrate with a bonding layer whose adhesion promoter layer has a surface resistance greater than 5 ohms / sq. US patent 2020 027 815 A1 deals with heat sinks for power modules.

[0007] The present invention therefore aims to provide improved support substrates which, in particular through their improved material selection with regard to the ceramic element, can achieve optimized heat dissipation.

[0008] This problem is solved by a carrier substrate according to claim 1 and a method according to claim 9. Further embodiments and configurations can be found in the dependent claims, the description and the figures.

[0009] According to a first aspect of the present invention, a support substrate for electrical components is provided which comprises the following: a heat sink, and a ceramic element, wherein the ceramic element is at least partially connected to the heat sink, wherein a solder-free bonding layer is formed between the heat sink and the ceramic element in the manufactured support substrate, and wherein an adhesion promoter layer of the bonding layer has a surface resistance greater than 5 ohms / sq, preferably greater than 10 ohms / sq and particularly preferably greater than 20 ohms / sq.

[0010] In contrast to the prior art carrier substrates comprising a heat sink and a ceramic element, the invention provides for a solder-free bonding layer formed between the ceramic element and the heat sink, which exhibits a comparatively high surface resistance. In other words, the bonding method used to connect to the ceramic element dispenses with a solder material, and the manufacturing process also results in a comparatively high surface resistance. This is particularly the case when the bonding to the ceramic element is achieved via hot isostatic pressing, preferably hot isostatic pressing in which an active metal layer is arranged between the ceramic element and the metal layer to be bonded. This creates a surface resistance that is determined primarily by the adhesion promoter layer.A defined bonding layer with a homogeneously distributed, comparatively small thickness (measured along a stacking direction).

[0011] A solder-free bonding layer is understood to be, in particular, a bonding layer that is essentially or exclusively attributable to active metal and contains no additional components attributable to a solder base material or an active metal-containing solder material. In other words, the bonding layer is preferably formed essentially exclusively by the adhesion promoter layer. During the bonding process, an active metal layer is transformed into the bonding layer or a part thereof.

[0012] Together with other parameters, such as the purity of the applied active metal layer and / or the roughness of the ceramic element, this contributes to a correspondingly high surface resistance. The elimination of a solder material when bonding the metal layer to the ceramic element (which in turn leads to the desired properties) has the advantage that the temperatures required during the heat sink bonding process do not impair the bond between the metal layer and the ceramic element. The heat sink is bonded in a subsequent step following the bonding of the metal layer to the ceramic element, at temperatures that would melt the solder material if bonded with solder. In contrast to bonding layers formed from solder, i.e.,Bonding layers containing solder material, exhibiting the required surface resistance, prove to be resistant to the temperatures present when the heat sink is bonded to the metal-ceramic substrate. This, in turn, advantageously makes it possible to provide support substrates whose ceramic elements were previously unsuitable because they could only be bonded to a metal layer via a solder system, particularly solder systems that cannot withstand the temperatures necessary for bonding the heat sink.

[0013] It is noted that the described bonding layer is located between the heat sink and the ceramic element and is specifically formed on or adjacent to the cooling side of the ceramic element. It is particularly emphasized that, due to the manufacturing process, at least one metal section may be formed between the heat sink (in the form in which it is provided for the bonding process) and the ceramic element, or between the installed heat sink and the ceramic element, viewed in the stacking direction. If this metal section is made of the same material as the heat sink used in the manufacturing process, a seamless transition between the heat sink and this metal section may occur.The person skilled in the art will therefore either attribute this metal section, which cannot be attributed to the originally used heat sink but is integrally connected to the provided heat sink in the manufactured state, to the heat sink in the manufactured state, or recognize that another metal layer or metal section may also be formed between the heat sink and the support substrate.

[0014] The substrate can be, for example, a printed circuit board (PCB) with conductive traces and / or connection pads on its component side, designed for connecting electrical components, particularly for forming electrical circuits. Alternatively, the substrate could be a cooling system used, for example, to cool a laser diode or a laser diode array. Preferably, the heat sink is formed by stacking at least one first and two metal layers, which are directly bonded together. Appropriate recesses within the first and second metal layers create a cooling channel system through which a cooling fluid, such as a liquid or gas, can be circulated during operation to dissipate heat from the substrate.It is also conceivable that the heat sink was manufactured differently and / or, for example, incorporates a fin structure. Regardless of the form in which it is provided, the cooling structure within the heat sink can be either open or closed.

[0015] To determine the surface resistance, the metal layer and, if applicable, a solder base layer are first removed from the manufactured substrate, for example, by etching. A four-point measurement is then taken at the top and bottom surfaces of the substrate, now free of at least one metal layer and the solder base layer. Specifically, the surface resistance of a material sample is defined as its resistance per unit area (a square of surface). It is common practice to express surface resistance in ohms per square (Ω / sq). The physical unit of surface resistance is the ohm.

[0016] Preferably, the substrate is designed as a printed circuit board (PCB) in which, in the manufactured state, the at least one metal layer bonded to the ceramic element is structured. For example, it is provided that, after the bonding step, structuring is also carried out, for example by laser treatment, etching, and / or mechanical processing, to create conductive traces and / or connections for electrical or electronic components. Preferably, a further metal layer, in particular a backside metallization, is provided on the ceramic element of a manufactured metal-ceramic substrate on the side opposite the metal layer. The backside metallization preferably serves to counteract deflection, and the heat sink, which is in turn bonded to the backside metallization, serves to effectively dissipate heat generated during the operation of electrical or electronic components.originates from electronic components that are attached to the circuit board or the metal-ceramic substrate.

[0017] Suitable materials for the metal layer in the metal-ceramic substrate or for the metal of the heat sink include copper, aluminum, molybdenum, tungsten, nickel, and / or their alloys such as CuZr, AlSi, or AlMgSi, as well as laminates such as CuW, CuMo, CuAl, and / or AlCu or MMC (metal matrix composite), such as CuW, CuM, or AlSiC. Furthermore, it is preferably provided that the at least one metal layer on the manufactured metal-ceramic substrate, particularly as component metallization, is surface-modified. Surface modification could, for example, involve sealing with a precious metal, particularly silver; and / or gold, or (electroless) nickel or ENIG ( "electroless nickel immersion gold" ) or edge sealing on the metallization to suppress crack formation or widening is conceivable.

[0018] Preferably, the thickness of the bonding layer or adhesion promoter layer, measured in the stacking direction and averaged over several measuring points within a predetermined area or in several areas parallel to the main extension plane, assumes a value that is less than 1000 nm, preferably less than 600 nm, and particularly preferably less than 350 nm. When referring to several areas, it is specifically meant that the metal layer is divided into areas of as equal size as possible, and that at least one, preferably several, thickness measurements are recorded in each of these areas dividing the metal layer. The thicknesses determined at different locations are then averaged arithmetically. In particular, the ceramic element has a material composition that cannot be bonded via a direct bonding method. Preferably, the ceramic element comprises silicon nitride.In particular, the ceramic element is designed to comprise more than 60 wt.%, preferably more than 80 wt.%, and most preferably more than 90 wt.% silicon nitride. Silicon nitride is particularly advantageous because it provides high thermal shock resistance and high flexural strength. Furthermore, the increased coefficient of thermal expansion reduces the formation of thermomechanical stresses. In addition, improved thermal conductivity enhances heat dissipation efficiency. It is particularly advantageous that the hot isostatic pressing used in the production of the metal-ceramic substrate allows the use of silicon nitride for such substrates.Finally, silicon nitride cannot be bonded to a metal layer, particularly a copper layer, using a direct bonding process, and the required solder materials have a melting point below 1000 °C, meaning that a direct bonding process for the heat sink would remelt the solder material. Therefore, the method, which is structurally reflected in the solder-free bonding layer with its increased surface resistance, also enables the use of silicon nitride in a suitable substrate. Furthermore, it is preferably provided that the ceramic element, in particular the silicon nitride ceramic, has a thermal conductivity greater than 90 W / mK, preferably greater than 110 W / mK, and particularly preferably greater than 120 W / mK. Moreover, it is conceivable that the ceramic element has a thickness of less than 300 µm, preferably less than 250 µm, and particularly preferably less than 200 µm.

[0019] Alternatively, the ceramic element comprises Al₂O₃, AIN, an HPSX ceramic (i.e., a ceramic with an Al₂O₃ matrix containing an x ​​percent ZrO₂, for example, Al₂O₃ with 9% ZrO₂ = HPS9 or Al₂O₃ with 25% ZrO₂ = HPS25), SiC, BeO, MgO, high-density MgO (> 90% of the theoretical density), or TSZ (tetragonally stabilized zirconia) as the ceramic material. It is also conceivable that the ceramic element is designed as a composite or hybrid ceramic, in which several ceramic layers, each differing in their material composition, are arranged one above the other and joined together to form a ceramic element in order to combine various desired properties. Preferably, the ceramic element is free of parylene.

[0020] According to the invention, the heat sink is connected to the ceramic element via a back-side metallization, and the adhesion promoter layer is formed between the back-side metallization and the ceramic element. The back-side metallization associated with the heat sink is bonded to the ceramic element in a preparatory manufacturing step, and the heat sink is only attached after the back-side metallization has been bonded. As a result, the back-side metallization section and the heat sink are subjected to different temperature treatments, which are reflected in different grain sizes. Consequently, these sections can be identified on the manufactured substrate.

[0021] Preferably, the grain size in the backside metallization differs from the grain size in the heat sink. Since, during the manufacturing process, the backside metallization can also be achieved during hot static pressing by fixing the metal layer to be bonded in a metal pouch, resulting in the bonding of both the metal layer and part of the metal pouch, structural differences can also be observed when metal materials of varying purity are used for the metal layer and the metal pouch.

[0022] For example, the bonding layer between the backside metallization and the ceramic element contains an active metal component, which is due to an active metal layer used in the manufacturing process.

[0023] Furthermore, it is conceivable that the backside metallization projects along a main plane of extension relative to the heat sink. Since the heat sink is attached to the backside metallization, which is already bonded to the ceramic element, and preferably does not extend over the entire backside of the ceramic element, the backside metallization can be made larger. This eliminates the need to partially remove the backside metallization, which is preferably bonded to the entire surface of the ceramic element. Moreover, this simplifies the positioning of the heat sink relative to the backside metallization during manufacturing, as an absolutely congruent arrangement is not required.

[0024] Preferably, the heat sink is formed, preferably, from at least a first metal layer and a second metal layer, which are joined to one another by means of a direct bonding method, in particular a solderless direct bonding method, wherein the first metal layer and / or the second metal layer have recesses that form a cooling channel in the manufactured support substrate. It is further conceivable that the cooling channel extends to the ceramic element and / or that a residual metal layer thickness is provided almost exclusively, and preferably without exception, between the ceramic element and the cooling channel. For the formation of a cooling channel extending to the support substrate, or...The cooling channel extending to the ceramic element is designed so that, in the manufacturing process, after the backside metallization is attached to the ceramic element, the backside metallization is structured to expose corresponding areas on the ceramic element before the cooling element is attached to the backside metallization.

[0025] Preferably, the distance between the side surfaces bounding a cooling channel is less than 0.5 mm, more preferably less than 0.4 mm, and particularly preferably less than 0.3 mm. Specifically, this refers to side surfaces spaced apart in a direction parallel to the main plane of extension. Such comparatively thin channels and distances between side surfaces can be produced, for example, by electrical discharge machining (EDM) or wire EDM in the respective metal layers. Such thin cooling channels are particularly advantageous because they ensure high efficiency in heat transfer and can also provide the most evenly distributed cooling and heat transfer possible on the cooling side.

[0026] Preferably, the thickness of the bonding layer or adhesion promoter layer, measured in the stacking direction and averaged over several measuring points within a predetermined area or in several areas parallel to the principal plane of extension, is less than 1000 nm, preferably less than 600 nm, and particularly preferably less than 350 nm. When referring to several areas, it is specifically meant that the metal layer is divided into areas of as equal size as possible, and that at least one, preferably several, thickness measurements are recorded in each of these areas dividing the at least one metal layer. The thicknesses thus determined at different locations are then averaged arithmetically.

[0027] Compared to the substrates known from the prior art, a comparatively thin bonding layer is thus formed between the at least one metal layer and the ceramic element. It is provided that, to determine the relevant thickness of the bonding layer, the measured thicknesses are averaged over a large number of measuring points located within a predetermined or defined area or areas.

[0028] In particular, it is provided that the adhesion promoter layer comprising an active metal has a substantially constant thickness. Specifically, the thickness measurements determined within the area(s) exhibit a distribution with a standard deviation of less than 0.2 µm, preferably less than 0.1 µm, and most preferably less than 0.05 µm. In particular, the physical and / or chemical vapor deposition of an active metal layer and the resulting bonding layer allow for a homogeneous and uniformly distributed thickness of the bonding layer, which in particular consists only of the adhesion promoter layer. The adhesion promoter layer can also have a constant thickness if it is formed in addition to the solder base material.

[0029] A further object of the present invention is a method for producing a substrate according to the invention comprising: bonding a backside metallization and preferably the component metallization to a ceramic element to form a metal-ceramic substrate by means of hot isostatic pressing, Providing a heat sink, in particular a heat sink with a cooling channel structure, and connecting the heat sink to the backside metallization by means of a direct bonding process, in particular a DCB process.

[0030] All the described advantages and properties of the carrier substrate can be applied analogously and transferred to the process and vice versa.

[0031] Preferably, during hot isostatic pressing, the metal container or a metal layer is exposed in a heating and pressure device to a gas pressure between 100 and 2000 bar, preferably between 150 and 1200 bar, and particularly preferably between 300 and 1000 bar, and a process temperature of 300 °C up to the melting temperature of a metal layer, in particular to a temperature below the melting temperature. It has been advantageously found that this makes it possible to bond a metal layer, for example, the component metallization and / or backside metallization, to the ceramic element without the temperatures required by a direct metal bonding process, such as a DCB or DAB process, and without a solder base material or solder material used in active soldering. Furthermore, the use of this method allows for...The use of a suitable gas pressure makes it possible to produce a metal-ceramic substrate with as few voids as possible, i.e., without gas inclusions between the metal layer and the ceramic element. In particular, process parameters are used that are mentioned in DE 2013 113 734 A1 and to which explicit reference is hereby made. Furthermore, it has been found that the bond produced in this way between the ceramic element and the metal layer can withstand temperatures greater than 1050 °C.

[0032] In particular, the present method is distinguished by the fact that the bonding of the backside metallization to the ceramic element is achieved through hot isostatic pressing rather than soldering. This results in a more robust bond between the component metallization or backside metallization on the one hand and the ceramic element on the other, especially with regard to the temperatures used when attaching the heat sink to the metal-ceramic substrate. Specifically, the heat sink is bonded to the backside metallization of the metal-ceramic substrate produced by hot isostatic pressing. Consequently, solder-based bonding methods are no longer necessary for using ceramic elements that cannot be bonded to a metal layer via a direct bonding process.Therefore, the described method also enables the attachment of such ceramic elements that could not previously be used for the formation of the described support substrates, since the solder materials used would otherwise be damaged or even destroyed in the attachment process of the heat sink.

[0033] Preferably, a recess is created in a first or second metal layer by etching, EDM, and / or milling, and at least the first and second metal layers are joined together to form the heat sink. This provides the heat sink for connection, with the heat sink providing a cooling channel system through which a cooling fluid can be guided to dissipate heat from the manufactured substrate.

[0034] Preferably, an active metal layer is arranged between the ceramic element and the component metallization or the backside metallization. This provides a particularly strong bond between the backside metallization and the ceramic element during hot isostatic pressing, with the active metal layer contributing significantly to the formation of the adhesion promoter layer or bonding layer. It is particularly desirable that the active metal layer becomes the adhesion promoter layer, especially the bonding layer, after the manufacturing process.

[0035] In particular, the use of a separately applied active metal layer makes it possible to design this layer to be comparatively thin, thus enabling the required relatively thin thicknesses of the bonding layer to be achieved, especially when averaged over various measurement values ​​within the defined area(s). Examples of active metals include titanium (Ti), zirconium (Zr), hafnium (Hf), chromium (Cr), niobium (Nb), cerium (Ce), tantalum (Ta), magnesium (Mg), lanthanum (La), and vanadium (V). It should be noted that the metals La, Ce, Ca, and Mg are easily oxidized. Furthermore, it should be noted that the elements Cr, Mo, and W are not classic active metals, but are suitable as a contact layer between Si₃N₄ and the at least one metal layer or the solder system or solder material, since they do not form intermetallic phases with the at least one metal layer, for example, copper, and do not exhibit edge solubility.

[0036] Preferably, the active metal is applied to the solder base material and / or the at least one metal layer and / or the ceramic element by chemical and / or physical vapor deposition, for example, by sputtering, in order to create comparatively thin active metal layers, which in turn result in a comparatively thin bonding layer, in particular a homogeneous and thin adhesion promoter layer. It is also conceivable to apply the active metal layer to the ceramic element and / or the metal layer for component metallization and / or backside metallization using a plasma in a vacuum and / or by vapor deposition. Electroplating the active metal layer is also possible. It is particularly preferred that the active metal layer be provided as a film.

[0037] The active metal layer can be produced, in particular, by gas-physical deposition, which allows for the creation of comparatively thin active metal layers that can also contribute to achieving the required surface resistance. Electroplating, electroless deposition, thermal deposition, or deposition by cold gas spraying are also conceivable.

[0038] Preferably, the ratio between the thickness of the active metal layer and the thickness of the first metal layer and / or second metal layer is between 0.0001 and 0.005, preferably between 0.005 and 0.003, and most preferably between 0.001 and 0.0015. Similarly, a comparatively thin active metal layer is advantageous for creating an effective bonding layer while keeping the consumption of active metal limited.

[0039] Preferably, the heat sink is bonded to the backside metallization by hot isostatic pressing after the metal-ceramic substrate has been formed. The metal-ceramic substrate is formed first to ensure that the backside metallization and the component metallization have the same or comparable thicknesses during the formation process. This prevents deflection due to thermomechanical stresses in the metal-ceramic substrate. In other words, a metal-ceramic substrate with component metallization and backside metallization is produced first to prevent deflection of the metal-ceramic substrate after or during the bonding process, which would otherwise result from the different thermomechanical expansion coefficients of the ceramic element and the metal layer.A heat sink can be more easily attached to such a largely flat and even metal-ceramic substrate, particularly via its top layer. Otherwise, the metal-ceramic substrate would have to be straightened, if at all possible, at considerable expense. Furthermore, it has been shown that metal layers thicker than 0.4 mm, when attached to the ceramic element, cause greater thermomechanical stresses after cooling, which persist even after cooling has ceased.

[0040] Further advantages and properties will become apparent from the following description of preferred embodiments of the invention with reference to the accompanying figures. These show: Fig. 1 Schematic representation of a support substrate according to an exemplary embodiment of the present invention Fig. 2:Schematic exploded view of a support substrate for an exemplary embodiment of the present invention, Fig. 3 schematic representation of a first metal layer for a heat sink made of Figure 1 , Fig. 4 schematic representation of a support substrate for a further exemplary embodiment of the present invention, Figs. 5 to 8: Detailed representation of the system consisting of the carrier substrate with a distribution structure

[0041] In Figure 1A support substrate 1 according to a preferred embodiment of the present invention is shown schematically. Such support substrates 1 are particularly designed to allow electrical components 4 to be connected to their component side 5. The support substrate 1 can, for example, be a printed circuit board with multiple connection pads, conductor tracks, and contact areas for implementing circuits. It is also conceivable that the support substrate 1 is essentially a cooling structure designed to selectively cool an electrical component, such as a laser diode or a laser diode array. To dissipate the heat generated during operation by the electrical components 4, the support substrate 1 includes a heat sink 20. This heat sink 20 is formed on a cooling surface 6 opposite the component side 5.Preferably, the heat sink 20 is a metallic structure formed by stacking at least one first metal layer 21 and a second metal layer 22, and then joining them together. Recesses formed in the first metal layer 21 and / or the second metal layer 22 allow for the creation of a channel system within the heat sink 20. A cooling fluid, such as a cooling gas or liquid, can be guided through this channel system to dissipate heat from the support substrate 1. To insulate the electrical components 4, particularly from the metallic heat sink 20, the support substrate 1 includes a ceramic element 71. Preferably, the support substrate 1 comprises a metal-ceramic substrate 70 and the heat sink 20, which are connected to form the support substrate 1.It is preferably provided that the metal-ceramic substrate 70 has a component metallization 72 on the component side of the ceramic element 71 and a back metallization 74 on a cooling side 6 of a ceramic element 71 of the metal-ceramic substrate 70 opposite the component side 5. The thicknesses of the component metallization 72 and the back metallization 74 are identical in order to compensate for thermomechanical stresses, particularly during the manufacturing step in which the component metallization 72 or the back metallization 74 is bonded to the ceramic element 71. These stresses arise due to the different coefficients of thermal expansion of the back metallization 74 / component metallization 72 on one side and the ceramic element 71 on the other. Without such compensation, the metal-ceramic substrate 70 tends to deflect.Therefore, it is preferably provided that the heat sink 20 is bonded to the metal-ceramic substrate 70, in particular to the back-side metallization 74 of the already manufactured metal-ceramic substrate 71. The bonding of a top side of the heat sink 20 to the back-side metallization 74 is preferably carried out by a direct bonding method, in particular by a solderless direct bonding method, which requires temperatures above 1,068 °C to generate a eutectic that forms between the back-side metallization 74 and the top layer of the heat sink 20. In the illustrated embodiment, the first metal layer 21 forms the top layer.

[0042] Preferably, the bonding process of the heat sink 20 to the metal-ceramic substrate 1 involves the first metal layer 21 and the second metal layer 22 already being bonded together. Alternatively, it is conceivable that the bonding between the first metal layer 21 and the second metal layer 22, and the bonding between the top layer of the heat sink 20 and the back-side metallization 74, are carried out simultaneously or at least overlapping in time. It is conceivable that the first metal layer 21 forms the top layer (see exemplary embodiment of the Figure 1 ) and / or the top layer is an additional layer that is bonded to a top surface of the heat sink 20. For example, it is conceivable that the top layer forms a continuous metal layer without cutouts.

[0043] The necessity of bonding the heat sink 20 to the metal-ceramic substrate 1, particularly to the back-side metallization 74, using a direct bonding process and therefore at the corresponding temperatures, means that the bond already formed between the ceramic element 71 and the component metallization 72 or the back-side metallization 74 must also withstand the manufacturing conditions for bonding the heat sink 20 to the metal-ceramic substrate 1. Otherwise, the manufacturing conditions required for bonding the heat sink 20 to the metal-ceramic substrate 1 would break or at least impair the bond between the back-side metallization 74 or component metallization 72 and the ceramic element 71. This is particularly the case if the bonding of the component metallization 72 or...For the backside metallization 74 to the ceramic element 71, solder materials must be used whose melting temperature is below 1000 °C.

[0044] However, this also means that ceramic elements 71 that can only be bonded to a metal layer (i.e., to component metallization 72 or backside metallization 74) via a solder material are excluded from use in the described support substrates 1. This applies, for example, to silicon nitride, which cannot be bonded to a metal using a direct bonding method. In order to also use such ceramic elements in the described support substrates, which would otherwise have to be excluded, the present invention provides for bonding the component metallization or the backside metallization to the ceramic element 71 by means of a solder-free bonding method, in particular a solder-free, hot isostatic pressing process.

[0045] In particular, it is provided that an active metal layer is arranged during the bonding process between the ceramic element 71 and the component metallization 72 and / or between the ceramic element 71 and the back-side metallization 74 during hot isostatic pressing. This layer is applied, for example, to the component side 5 or the cooling side 6 of the ceramic element 71 and / or to the component metallization 72 and / or back-side metallization 74 before hot isostatic pressing, for example by a gaseous deposition process, such as a sputtering process, or an electrochemical process. It has been shown that bonding between the ceramic element 71 and the component metallization 72 or back-side metallization 74 is possible, especially for those ceramic elements 71 that are not accessible for a direct metal bonding process.At the same time, it has been shown that the connection can withstand the temperatures required for the direct bonding process of the heat sink 20 to the metal-ceramic substrate 70. This is primarily because the bond between the ceramic element and the metal layer can be made without solder or solder base material in this case.

[0046] It is emphasized once again that the active metal layer is an active metal layer whose active metal content is greater than 15 wt.%, preferably greater than 30 wt.%, and particularly preferably greater than 70 wt.%. It is therefore not an active metal-containing solder layer, as is otherwise customary when bonding a component metallization 72 to a ceramic element 71. This allows, for example, the use of ceramic elements 71 that are particularly advantageous due to their high thermal conductivity for forming a corresponding support substrate 1. This applies in particular to support substrates 1 whose ceramic element 71 comprises silicon nitrides or is composed of more than 80% silicon nitride.

[0047] The described bonding method is preferably characterized by the fact that the bonding layer A bonding layer exists between the ceramic element 71 and the component metallization 72 and / or between the ceramic element 71 and the backside metallization 74, the surface resistance of which is greater than 5 ohms / sq, preferably greater than 15 ohms / sq, and particularly preferably greater than 20 ohms / sq. The described surface resistance is, in particular, the result of a uniformly applied active metal layer and bonding via hot isostatic pressing. Bonding layers produced using a solder material with an active metal component, on the other hand, exhibit a lower surface resistance. In particular, it is provided that only an adhesion promoter layer is formed as the bonding layer between the backside metallization 74 or component metallization 72 and the ceramic element 71, i.e., no sub-areas attributable to a solder material are identifiable in the manufactured carrier substrate 1.

[0048] Since the bonding process of the backside metallization 74 to the top layer of the heat sink 20, using a direct bonding method, does not result in a visible boundary / transition zone at the manufactured interface between the backside metallization 74 and the top surface of the heat sink 20, the backside metallization 74 is, in the manufactured state, part of the heat sink 20. Thus, the surface resistance between the ceramic element 71 and the heat sink 20 is formed, to which, in the manufactured state, the original backside metallization 74 of the metal-ceramic substrate 1 is attributed. Therefore, the metal-ceramic substrates produced according to the inventive method can be identified by the corresponding surface resistance formed between the ceramic element 71 and the heat sink 20. Furthermore, it is specifically intended that a bonding layer forms, which is exclusively an adhesion promoter layer and is not attributable to a solder material.

[0049] Furthermore, in Figure 1 A meandering or loop-shaped channel system is formed into which, for example, a cooling fluid can be introduced via an inlet E. This fluid is guided through the cooling element 20 in a looping or meandering motion and finally exits the cooling element 20 through an outlet opening A. Further examples of cooling elements that could be provided for certain embodiments of the invention and that can be connected to a metal-ceramic substrate 70, in particular its backside metallization, are shown in the following figures.

[0050] In Figure 2Figure 1 schematically shows an exploded view of a heat sink 1 according to a first preferred embodiment for an object according to the present invention. In particular, the heat sink 1 is designed for cooling an electronic or electrical component (not shown), especially a semiconductor element and, more preferably, a laser diode. For cooling the electrical or electronic component, the manufactured heat sink 1 forms a cooling fluid channel system through which a cooling fluid can flow during operation, so that the cooling fluid can absorb and dissipate heat emitted by the electronic or electrical component during operation.

[0051] For this purpose, the cooling element 1, particularly the cooling fluid channel system, preferably includes an inlet area and an outlet area (not shown), wherein the cooling fluid is introduced via the inlet area and discharged again via the outlet area. Preferably, the cooling fluid channel system is designed such that the cooling fluid 1 passes through a fin structure 25 at the transition from the inlet area to the outlet area, which in particular projects into the cooling fluid channel system. The fin structure 25 preferably consists of rib-like elements 7 that project into the cooling fluid channel system in order to provide the largest possible contact area for the fluid, thus enabling effective heat transfer from the rib-like element 7 or the wall of the cooling channel system to the fluid.

[0052] Preferably, the heat sink 1 comprises at least one first metal layer 11, at least one second metal layer 12 and / or at least one third metal layer 13. To form the cooling fluid channel system, the at least one first metal layer 11, the at least second metal layer 12 and / or the at least third metal layer 13 are structured by at least one recess 21, 22 such that they form the cooling fluid channel system by being stacked on top of each other or placed one above the other along the stacking direction S.

[0053] It is specifically provided that the at least one first metal layer 11, the at least one second metal layer 12, and / or the at least one third metal layer 13 are each structured differently or equipped with differently oriented recesses 21, 22. In particular, it is provided that the at least one first metal layer 11, the at least one second metal layer 12, and / or the at least one third metal layer 13 form at least one first part 21 in the at least one recess 21, 22, which has the web-like elements 7 that extend, in particular, in a principal extension plane HSE perpendicular to the stacking direction S.In addition to the first part 21 of the at least one recess 21, 22 in the at least one first metal layer 11, it is preferably provided that a second part 22 of the at least one recess 21, 22 in the at least one first metal layer 11 is provided for supplying or discharging the cooling fluid into or out of the first part 21 or forms a part of the supply area and / or discharge area.

[0054] The heat sink 1 is preferably bounded in the stacking direction S by an upper cover layer 15 and a lower cover layer 14, wherein the first, at least one, metal layer 11, the at least one second metal layer 12, and / or the at least one third metal layer 13 are arranged between the lower cover layer 14 and the upper cover layer 15 in the stacking direction S. In particular, the formation of the at least one first metal layer 11, the at least one second metal layer 12, and / or the at least one third metal layer 13 is arranged in a sandwich-like configuration between the upper cover layer 15 and the lower cover layer 14. In addition to the at least one recess 21, 22, which is composed of the first part 21 and the second part 22, it is preferably provided that the heat sink 1 or the at least one first metal layer 11 has a further recess 24, which is not part of the cooling fluid channel system with the fin structure 25.Furthermore, it is preferably provided that a connection surface 30 is provided on the upper cover layer 15 and / or the lower cover layer 14. In particular, the electrical or electronic component is connected to this connection surface 30, especially when viewed in the stacking direction S, above or below the fin structure 25, which preferably extends in a direction perpendicular to the stacking direction S. In other words, the fin structure 25, especially its rib-like elements 7, extends below the connection surface 30 and preferably parallel to it. By appropriately arranging the fin structure 25 with its rib-like elements 7 above or below the connection surface 30, the electrical or electronic component can be effectively cooled by means of the fin structure 25.

[0055] In Figure 3is a schematic representation of at least one first metal layer 11, which is shown, for example, in Figure 1The fin structure 25 is installed in the illustrated embodiment. It is formed from rib-like elements 7 that extend to different distances in the main extension plane HSE. In particular, the length of the rib-like elements 7 increases towards a central axis M of the at least one first metal layer 11. This advantageously maximizes the cooling effect, especially in the central region of the contact surface 30. Furthermore, it is conceivable that the rib-like elements run parallel and / or obliquely to the central axis M. Preferably, the shape of the rib-like elements 7, in particular their length and / or inclination relative to the central axis M along the main extension plane HSE, is determined or specified by the corresponding cooling requirements of the electrical or electronic component.

[0056] To achieve the smallest possible distance A1 between two adjacent rib-like elements 7, it is provided, for example, that the first part 21 of the at least one recess 21, 22 in the at least one first metal layer 11 is produced by electrical discharge machining (EDM), in particular by spark EDM. Specifically, this involves production by wire EDM.

[0057] Furthermore, it is provided that a second part 22 of the at least one recess 21, 22 is formed by etching. Preferably, the etching is carried out particularly in large areas of the second part 22 of the recess 21, 22, i.e., in the subsequent inlet and / or outlet areas designed for supplying and discharging the cooling fluid. In contrast, it is particularly provided that electrical discharge machining (EDM) is used for the fine-structured forming of the recess 21, 22, i.e., the first part 21 of the recess 21, 22. It has been found that this allows comparatively very small gaps between the rib-like elements 7 to be produced without having to rely on several first metal layers 11 with etched first parts 21 of the at least one recess 21, 22, which would have to be stacked on top of each other to achieve the smallest possible gap between two rib-like elements 7.Preferably, the distance A1 between opposing side walls between two rib-like elements 7 is less than 0.4 mm, more preferably less than 0.3 mm, and particularly preferably less than 0.2 mm. This allows as many rib-like elements 7 as possible to be integrated into the fin structure 25. Accordingly, it is possible to increase the cooling effect, since the contact area between the cooling fluid and the surface of the cooling fluid channel system can be increased accordingly.

[0058] Preferably, the at least one first metal layer 11, the at least one second metal layer 12, the at least one third metal layer 13, the upper cover layer 15 and / or the lower cover layer 14 have a thickness, measured in the stacking direction S, of between 0.2 and 0.7 mm, preferably between 0.35 and 0.6 mm, and particularly preferably between 0.3 and 0.4 mm. Preferably, the at least one first metal layer 11, the at least one second metal layer 12 and / or the at least one third metal layer 13 each have the same thickness.Furthermore, it is preferably provided that the at least one first metal layer 11, the at least one second metal layer 12, and / or the at least one third metal layer 13 are formed into an integral cooling fluid channel system by means of a sintering process in which the microstructures of the at least one first metal layer 11, the at least one second metal layer 12, and / or the at least one third metal layer 13 merge or fuse together through a corresponding heat treatment. It is further provided that the upper cover layer 15 and / or the lower cover layer 14 each have at least one recess 21, 22, and / or a further recess 24, wherein the upper cover layer 15 and / or the lower cover layer 14 are preferably free of rib-like elements 7 or components of a subsequent fin structure 25. The further recesses 24 preferably serve for fastening or fixing the cooling element 1.

[0059] In Figure 4A metal-ceramic substrate 70 according to a further exemplary embodiment of the present invention is shown schematically. Such metal-ceramic substrates 1 preferably serve as carriers for electronic or electrical components 4, which can be bonded to the metal-ceramic substrate 70. Essential components of such a metal-ceramic substrate 1 are a ceramic element 71 extending along a principal extension plane HSE and a component metallization 72 bonded to the ceramic layer 71. The ceramic element 71 is made of at least one material comprising a ceramic. The component metallization 72 and the ceramic element are arranged one above the other along a stacking direction S extending perpendicular to the principal extension plane HSE and are metallurgically bonded to one another.In the manufactured state, the component metallization 72 on one component side 5 of the metal-ceramic substrate 70 is structured to form conductive tracks or connection points for the electrical components 4. In the illustrated embodiment, the metal-ceramic substrate 70 comprises a secondary layer 73 and a metallic intermediate layer 75 arranged between the ceramic element 71 and the secondary layer 73. The ceramic element 71, the metallic intermediate layer 75, and the secondary layer 73 are arranged one above the other along the stacking direction S. Furthermore, it is provided that the metallic intermediate layer 75 is thicker than the ceramic element 71 and / or the secondary layer 73. Preferably, the metallic intermediate layer 75 is thicker than 1 mm, more preferably thicker than 1.5 mm, and particularly preferably thicker than 2.5 mm.However, it is also conceivable that the metal-ceramic substrate is formed using only a ceramic element 71 with a component metallization 72 and a backside metallization. While the ceramic element 71 is preferably made of ceramic and designed to provide sufficient insulation strength and to stiffen the metal-ceramic substrate 1, the secondary layer 73 can, for example, also be made of tungsten or molybdenum, since a high degree of insulation strength is not required here. This reduces material costs. Alternatively, the secondary layer 73 is also made of a material comprising a ceramic.

[0060] A metallic heat sink 20 is provided on one of the cooling sides 6 of the metal-ceramic substrate 1, opposite the component side 5. The metallic heat sink 20 is preferably directly connected to the secondary layer 73. However, it is also conceivable that the heat sink 20 is directly connected to a back-side metallization of the metal-ceramic substrate 1 or to the ceramic element 71 of the metal-ceramic substrate 1. This prevents an interface that would otherwise form with a corresponding bonding material from negatively affecting the thermal conductivity and thus restricting heat transfer from component side 5 to the cooling side 6.

[0061] For example, the cooling structure 20 is directly bonded to the secondary layer 73, the backside metallization, and / or the ceramic element 71 via an AMB process, a DCB (direct copper bonding) or DAB (direct aluminum bonding) process. In particular, a multitude of fluid channels 30 are integrated into the metallic cooling structure 20. For clarity, Figure 1 Only one of these fluid channels 30 is shown as an example. The fluid channels 30 serve to guide a fluid, in particular a cooling fluid, within the metallic cooling structure 20. The fluid is supplied to the cooling structure 20 via a distribution structure 40 and discharged again via the distribution structure 40. Preferably, the distribution structure 40 has an inlet section 41 and a discharge section 42 for this purpose.

[0062] In particular, the fluid channel 30 has an inlet opening 31 and an outlet opening 32 spaced apart from the inlet opening 31. The inlet opening 31 and the outlet opening 32 are part of an outer surface A of the cooling structure 20 facing the distribution structure 40. Specifically, the inlet part 31 of the distribution structure 40 borders the inlet opening 31 and the outlet part 42 borders the outlet opening 32.

[0063] In the Figures 4 to 6 Only the metallic cooling structure 20 and the distribution structure 40 are shown in perspective views and in the Figure 7 and 8The figures are shown in two different side views. Instead of the entire cooling structure 20, the figures depict the multiple fluid channels 30 of the cooling structure 20. In other words, the fluid channels 30 are shown here without the metallic body in which they are embedded. Viewed in the stacking direction S, a feed structure 50 is connected to the underside of the distribution structure 40. Thus, viewed in the stacking direction S, the distribution structure is arranged between the cooling structure and the feed structure. Such a feed structure 50 is preferably designed to define a first main flow direction HS1. For example, the feed structure is channel-shaped. Furthermore, the feed structure 50 includes at least one inlet and one outlet (not shown here) to which a fluid circuit can be connected, i.e., a cooling fluid supply and a cooling fluid discharge.The distribution structure 40 is preferably configured such that it deflects or introduces the fluid from a flow along the first main flow direction HS1 into the cooling structure 20. For the sake of clarity, the figures show only a single row of fluid channels 30. Preferably, several rows are arranged side by side or one behind the other in a direction perpendicular to the row direction RR and parallel to the main extension plane HSE, and each of these rows is supplied with the fluid via a corresponding distribution structure 40, for example, a single distribution structure. Preferably, these multiple rows extend completely over the outer surface A of the cooling structure 20 facing the distribution structure.

[0064] In particular, it is provided that several fluid channels 30 are arranged side by side. Specifically, in the illustrated embodiment, the fluid channels 30 are arranged along a row which, in the illustrated embodiment, runs essentially perpendicular to the first main flow direction HS1. It is also conceivable that the row runs along a row direction RR which is inclined at an angle between 0° and 90° relative to the first main flow direction HS1. Preferably, the angle is less than 45°.

[0065] In the Figures 4 to 8In the illustrated embodiment, the distribution structure 40 is designed to deflect the fluid such that at least a portion of the fluid is first deflected from a first main flow direction HS1 into a transverse direction Q parallel to the series direction before being directed into the inlet openings 31. Additionally, the fluid is deflected upwards towards the cooling structure 20 in the direction of the inlet opening 31. This allows the distribution structure to supply several inlet openings 31 of different fluid channels 30 with a fluid at the same temperature. For this purpose, the inlet section 41 is designed as a wall-like structure, which in the illustrated embodiment runs essentially parallel to the series direction RR. Preferably, the feed structure 50 supplies the fluid to only a portion of the distribution structure 40.In the illustrated embodiment, along the first main flow direction HS1, essentially a first part, in particular a left half, of the inlet section 41 is supplied with the cooling fluid. However, the entire series of fluid channels 30 is supplied with the fluid by means of the distribution structure 40. Preferably, the inlet section comprises a ramp-like structure that is inclined in the series direction RR, in particular with respect to the main extension plane HSE.

[0066] After passing through the fluid channels 30, the fluid exits the cooling structure 20 via the outlet openings 32 and is directed into the discharge section 42 of the distribution structure. The discharge section 42 of the distribution structure 40 is also designed as a wall-like structure that runs essentially parallel to the series direction RR. In particular, the discharge section 42 is designed to collect the fluid exiting the outlet openings and redirect it back into the feed structure 50 in a second main flow direction HS2. For example, the discharge section 42 includes a ramp-like structure inclined in the series direction RR, in particular inclined in the opposite direction to the ramp-like structure in the inlet section 41 of the distribution structure 40. Furthermore, the first main flow direction HS1 and the second main flow direction HS2 are offset from each other.In other words, after leaving the distribution structure 40, the flow of the fluid is laterally displaced compared to the flow when approaching the distribution structure 40.

[0067] In the illustrated embodiment, the inlet part 41 of the distribution structure 40 is arranged upstream of the outlet part 42 of the distribution structure 40 when viewed along the first main flow direction HS1. However, it is also conceivable that the outlet part 42 is arranged upstream of the inlet part 31 of the distribution structure 40 when viewed along the first main flow direction HS1.

[0068] The individual fluid channels 30 are preferably U-shaped, with the U-shaped fluid channel 30 having two leg sections 34 extending substantially perpendicular to the main extension plane HSE and a transverse section 33 connecting the two leg sections 34. In particular, the transverse section 33 serves to deflect the fluid and, in the installed state, is closest to the secondary layer 13 or the ceramic layer 11. Preferably, the distance between the transverse section 33 and a ceramic layer 11 or secondary layer 13 adjacent to the cooling structure 20 is between 0.2 and 1.5 mm, more preferably between 0.4 and 1 mm, and most preferably between 0.6 and 0.8 mm. Preferably, the fluid channels 30, and in particular their leg sections 34, are designed such that the fluid is swirled within the fluid channels 30.For this purpose, it is provided, for example, that an opening cross-section Q1, Q2 of the leg sections 34, extending parallel to the main extension plane HSE, is laterally displaced along a flow direction of the fluid within the fluid channel 30, particularly within the leg section 34. The leg section 34 comprises a first subsection T1 with a first opening cross-section Q1 and a second subsection T2 with a second opening cross-section Q2, wherein the first opening cross-section Q1 is offset from the second opening cross-section Q2 by an offset distance V in a direction parallel to the main extension plane HSE. Preferably, the first opening cross-section Q1 and the second opening cross-section Q2 are of the same size. However, it is also conceivable that the first opening cross-section differs from the second opening cross-section.In particular, the first subsection T1 and the second subsection T2 are each assigned metal layers, which are stacked on top of each other, for example, during manufacturing. The individual metal layers can be of the same thickness or differ in thickness. It is also conceivable, for example, that the thickness of the individual layers decreases and / or increases towards the side of the component.

[0069] In particular, it is provided that the first opening cross-section Q1 and the second opening cross-section Q2 are offset from each other in two directions: one parallel to the first main flow direction HS1 and the other parallel to the series direction RR, i.e., in two directions that are not parallel to each other. Preferably, the ratio of an overlap area, in which the first opening cross-section Q1 and the second opening cross-section Q2 are arranged one above the other in the stacking direction S, to the first opening cross-section Q1 and the second opening cross-section Q2 is between 0.5 and 0.9, preferably between 0.5 and 0.8, and most preferably between 0.5 and 0.7. It is particularly conceivable that the opening cross-section of the inlet opening and / or the outlet opening is larger than the first opening cross-section and / or the second opening cross-section.This allows a funnel-shaped inlet and outlet area to be formed for the fluid channel.

[0070] It is also conceivable that the first opening cross-section Q1 and the second opening cross-section Q2 are of different sizes. Preferably, the first and second opening cross-sections are designed such that they form an essentially spiral path for the fluid channel 30. The fluid channels 30 can be realized, for example, by layering metal layers, i.e., at least a first metal layer 11 and a second metal layer 12, with corresponding openings, or by a 3D printing process. Furthermore, it is provided that the inlet opening 13 has a first opening cross-section whose diameter and / or edge length has a value between 0.1 mm and 2.5 mm, preferably between 0.5 mm and 1.5 mm, and preferably substantially 1 mm. It is preferably provided that the first opening cross-section Q1 and the second opening cross-section Q2 do not change within the leg regions 34 of the fluid channel 30.

[0071] Furthermore, it is provided that a further distance A2 between two adjacent leg sections, preferably of the same fluid channel, assumes a value between 0.1 mm and 5 mm, preferably between 0.2 mm and 2 mm, and particularly preferably substantially 1.5 mm. The further distance A2 is measured between two centers of the first cross-sectional opening Q1 or second cross-sectional openings Q2 at the same height, viewed in the stacking direction S.

[0072] In addition to the heat sinks 20 and support substrates 1 described above, other heat sinks 20 with different geometries are also conceivable. For example, it is even possible for the heat sink 20 to be formed from a substantially flat, unstructured metal body 20. However, a structured heat sink 20 is preferred, which, for example, is open and / or at least partially closed or completely closed on its cooling side. Preferably, microchannels are formed in the manufactured heat sink 20, through which a cooling fluid, preferably a coolant, can then be guided during operation. This allows the heat generated, for example, on the component side 4 or caused by a laser diode mounted on the connection surface 30, to be dissipated.To prevent the metal heat sink 20 from corroding over time, it is designed that the heat sink 20 has at least a partial corrosion protection layer 30. Otherwise, such corrosion would cause the heat sink 20 to either leak and / or clog the microchannels or cooling areas. 1 Support element 4 Electrical component 5 Component side 6 Cooling side 7 Web-like element 11 First metal layer 12 Second metal layer 13 Third metal layer 14 Lower cover layer 15 Upper cover layer 21 First part 22 Second part 24 Further recess 25 Fin structure 30 Connection surface 33 Transverse area 34 Leg area 40 Distribution structure 41 Inlet part 42 Discharge part 50 Feed structure 70 Metal-ceramic substrate 71 Ceramic element 72 Component metallization 75 Intermediate layer 73 Secondary layer A1 Spacing A2 Further spacing S Stack direction T1 First section T2 Second section Q Transverse direction Q1 First cross-sectional opening Q2 Second cross-sectional opening M Central axis D Thickness V Offset spacing RR Series direction HS1 First main flow direction HS2 Second Main flow direction HSE Main extent plane

Claims

1. A carrier substrate (1) for electrical components (4), comprising: - a heat sink (20), and - a ceramic element (71), wherein the ceramic element (71) is bonded to the heat sink (20) at least in sections, wherein a bonding layer free of solder material is formed in the manufactured carrier substrate (1) between the heat sink (20) and the ceramic element (71), and wherein an adhesion agent layer of the bonding layer has a sheet resistance which is greater than 5 ohm / sq, more preferably greater than 10 ohm / sq and most preferably greater than 20 ohm / sq, wherein the heat sink (20) is bonded to the ceramic element (71) via a backside metallization (74) and the adhesion agent layer is formed between the backside metallization (74) and the ceramic element (71).

2. The carrier substrate (1) according to claim 1, wherein the ceramic element (71) comprises a material composition which cannot be bonded via a direct bonding method.

3. The carrier substrate (1) according to any one of the preceding claims, wherein the ceramic element comprises Si3N4.

4. The carrier substrate (1) according to any one of the preceding claims, wherein a grain size in the backside metallization (74) is different from a grain size in the heat sink (20).

5. The carrier substrate (1) according to any one of the preceding claims, wherein the ceramic element has a thickness which is less than 300 µm, more preferably less than 250 µm and most preferably less than 200 µm.

6. The carrier substrate (1) according to any one of the preceding claims, wherein the heat sink is formed from at least a first metal layer (21) and a second metal layer (22) which are joined on top of one another by means of a direct bonding method, in particular a solderless direct bonding method, wherein the first metal layer (21) and / or the second metal layer (22) have recesses which form a cooling channel in the manufactured carrier substrate (1).

7. The carrier substrate (1) according to any one of the preceding claims, a thickness of the bonding layer or adhesion agent layer, measured in the stacking direction (S), averaged over a plurality of measuring points within a predetermined area or in a plurality of areas which course or run parallel to the main extension plane, has a value which is less than 1000 nm, more preferably less than 600 nm and most preferably less than 350 nm.

8. The method of manufacturing a carrier substrate (1) according to any one of the preceding claims, comprising: - bonding a backside metallization (74) and preferably a component metallization (72) to a ceramic element (71) by means of hot isostatic pressing for forming a metal-ceramic substrate, - providing a heat sink (20), in particular a heat sink (20) with a cooling channel structure, and - bonding the heat sink (20) to the backside metallization (74) by means of a direct bonding method, in particular a DCB (direct copper bonding) method.

9. The method according to any one of the preceding claims, wherein a void in a first metal layer (21) and / or the second metal layer (22) is realized by means of etching, eroding and / or milling and at least the first metal layer (21) and the second metal layer (22) are joined together by means of a DCB process to form the heat sink (20), wherein the bonding of the at least first metal layer (21) and / or second metal layer (22) to the backside metallization and the bonding of the first metal layer (21) and the second metal layer (22) to form the heat sink (20) are performed at least partially simultaneously.

10. The method according to any one of the preceding claims, wherein an active metal layer for forming the bonding layer is arranged between the ceramic element (71) on the one hand and the component metallization (72) and / or the backside metallization (74) on the other hand.

11. The method according to claim 10, wherein a ratio between a thickness of the active metal layer and a thickness of the first metal layer (21) and / or second metal layer (22) has a value between 0.0001 and 0.005.

12. The method according to any one of the preceding claims, wherein the metal-ceramic substrate (70) is produced by hot isostatic pressing and thereafter the bonding of the heat sink (20) to the backside metallization (74) is realized.