WO2020152284A1

WO2020152284A1 - Process for manufacturing a ceramic-matrix composite part, and corresponding composite part and electrical component

Info

Publication number: WO2020152284A1
Application number: PCT/EP2020/051659
Authority: WO
Inventors: Driss KENFAUI; Marie-Laure LOCATELLI; Zarel VALDEZ NAVA; Lionel LAUDEBAT; Vincent Bley; Pascal Dufour; Sophie Guillemet; Christophe TENAILLEAU
Original assignee: Universite Toulouse Iii - Paul Sabatier; Centre National De La Recherche Scientifique; Institut National Universitaire Jean-François Champollion
Priority date: 2019-01-23
Filing date: 2020-01-23
Publication date: 2020-07-30
Also published as: FR3091867B1; FR3091867A1

Abstract

The present invention relates to a process for manufacturing a composite part (11) comprising: - a substrate (111) composed of a first material (A), said first material (A) being a ceramic or a ceramic-matrix composite; and - a plurality of three-dimensional volumes (112) arranged in said substrate (111), each of said three-dimensional volumes (112) being composed of a second material (B) which is a ceramic or a ceramic-matrix composite having a predefined electrical conductivity, permittivity and thermal conductivity, said three-dimensional volumes (112) being composed of one and the same ceramic or of one and the same ceramic-matrix composite or else of different ceramics or ceramic-matrix composites, the method comprising the following steps: producing (S30) a first preform (P1) composed of the first material (A) and intended to form the substrate (111) and obtaining a plurality of hollow impressions (EM) on a face of said first preform (P1), each of said hollow impressions (EM) being configured to delineate one of said three-dimensional volumes (112), completely or partially filling (S31) each of said hollow impressions (EM) with the second material (B) so as to produce second preforms (P2) each forming one of said three-dimensional volumes (112), joining (S33), by means of sintering, the first preform (P1) and the second preforms (P2) so as to obtain the composite part (11).

Description

Manufacturing process of a composite part with ceramic matrix, composite part, and corresponding electrical component

Field of the invention

The present invention relates to a method of manufacturing a part made of composite material.

One field of application of the invention is the manufacture of electrical components. The invention applies more generally, but not exclusively, to electronic, power electronics or electrical engineering systems intended for the conversion, distribution or transport of high voltage electrical energy (such as power converters, high voltage transformers, isolating gas switches, high voltage connectors, elements for high voltage transmission, high voltage cables whether in DC ("Direct Current") or AC ("Alternative Current" »), The bus-bars, for example).

The term "electrical component" in this document must be interpreted in the broad sense and can correspond equally well to an electrical or electronic module, an electrical or electronic circuit, a printed circuit of the PCB type, an electrical or electronic card, an electronic component, an electrical connector, an electrical cable, etc. More generally, the invention can be applied to any element with electrical or electronic functionality provided with electrical contacts intended to be surrounded by an electrical insulator.

Prior art

Industrial ceramics (called technical) constitute a set of advanced materials, used in a very wide variety of applications ranging from the field of electronics to that of space. They have multiple and varied functional properties (electrical, magnetic, thermal, optical, etc.), good mechanical properties and / or excellent thermal and chemical stability, associated with low density. The local control of the properties, in one or more parts, of a ceramic is crucial to give it the capacity to fulfill more technological functions and / or to reinforce its mechanical resistance.

Such an approach is sought for new ranges of applications in search of materials that are more functional, efficient, reliable, stable, light and / or resistant to temperature rise and under severe environments.

It therefore requires the development of composite parts with a ceramic matrix. Current composite parts are obtained by assembling, at a planar and two-dimensional (2D) interface, two ceramics with distinct chemical compositions.

Mention may be made, by way of example, of those developed for solar cell applications. These parts are generally manufactured by co-sintering two layers of powders, previously superimposed, using the techniques of forming by hot pressing. One of the two elaborate ceramics serves as a substrate.

While this approach makes it possible to confer new surface properties on this substrate, it does not allow a modification in volume thereof.

Even in the case of additive manufacturing techniques such as the selective laser sintering process ("Selective Laser Sintering" or SLS) or the selective laser melting process ("Selective Laser Melting" in English), which are techniques for the development of complex three-dimensional shapes and objects, the materials produced exhibit properties much lower than those obtained with conventional consolidation techniques (sintering without or under load), and the complexity of the laser-material interaction limits their optimal application to one material at a time.

This consequently precludes their use for many applications requiring the control of functional properties by volume.

It therefore appears necessary to propose an innovative solution making it possible to modify in volume a ceramic substrate so as to locally control its functional properties.

It will be recalled that certain ceramics are excellent electrical insulators and are used, for example, as an insulating substrate for power electronics applications. Electrical insulation is a critical element in devices operating under high voltage. Since the life of an electrical component is often linked to the life of its insulation, research has been carried out in recent years to understand the causes of degradation and aging of materials used as solid electrical insulators, and in particular in power electronics modules. The latter, operating at high voltage, are in fact subjected to strong electrical stresses, which can lead to the appearance of electric field reinforcements in the insulating material (ie singular zones near the conductive elements around which the electric field is more intense). These field reinforcements are - on condition that certain voltage levels are reached - sources of partial discharges in the volume of the insulators, deteriorating the component as it is used, or even electrical breakdown.

This is all the more remarkable given that the insulating materials used in current power components endure stresses for which they were not necessarily designed. Indeed, the rise in operating voltage combined with the sharp increase in the integration of electronics in on-board systems in particular, leads to an increase in power density. As a result, the level of electrical stresses to be endured for the insulating materials is multiplied.

One solution to guarantee better voltage withstand would consist in oversizing the constituent elements of the electrical components, but this approach is obviously not compatible with a search for system integration.

A component frequently used in energy conversion systems for rail transport, for example, is the power module with IGBT chips (for "Insulated Gâte Bipolar Transistor" or insulated gate bipolar transistor), as shown in Figure 1. Its structure consists of a stack of different elements. One or more IGBT chips 1 are soldered onto an insulating substrate 2 comprising an electrically insulating layer 21 (based on a ceramic material) which is covered on its lower and upper faces with a metallic electrical contact 22 and 23 (metallization). The substrate 2 is a metallized ceramic substrate (Type AI203, Si3N4, AIN) called DBC (for “Direct Bonding Copper”) or AMB (for “Active Metal Brazing”) due to the process by which the metallization is carried out. In the most common case, this substrate is placed on a support 5 made of copper or AlSiC. The different elements of the module are assembled and covered with a layer of electrically insulating material 3 (encapsulant, most often a silicone gel or an epoxy resin), then enclosed in a plastic case 6. The module is then fixed to a system of cooling the lower face of the module (not shown), in order to dissipate the heat produced by the chips in operation, then connected by conductors to the rest of the electrical circuit (actuators, sources, etc.).

One of the main causes of failure from an electrical point of view of this type of component is the breakdown of the electrical insulation at the ceramic / metal / encapsulation insulator triple point, as shown in Figure 2 (triple point denoted PT) corresponding to the interface of three media of different permittivities, and at the level of the tip of the electrical contact 22 (referenced P) at the metal / insulator interface (in this case we speak of “peak effect”) . Indeed, the strengthening of the electric field localized in the insulators in the vicinity of this sensitive point exists and can lead to the formation of partial electric discharges, sometimes manifesting in the form of electrical trees, the repetition of which degrades the encapsulation material 3, leading to premature aging of the module, as well as reliability problems, or a limitation of the voltage withstand. The field reinforcement is also strongly linked to the geometry of the edges of the contacts which are obtained after etching and amplifies the non-uniformity of the electric field in the insulating materials.

With a view to increasing voltage and / or integrating power electronics while maintaining high voltage constraints, a known solution, described in patent document WO 2015/074431, and illustrated in FIG. 3, consists in depositing a thin layer of semi-resistive varnish 4 based on hydrogenated amorphous silicon on the ceramic layer 21 at the interface with the encapsulation material 3 from the upper electrical contact 22 to the lower electrical contact 23. This semi-resistive varnish layer 4 makes it possible to reduce the risk of partial discharges located around the electrical contacts. However, such a solution requires the use of plasma deposition equipment which is expensive and difficult to read. In addition, the masking of certain elements sensitive to within the structure is relatively complex in view of the variation of the structural elements that can compose it.

Another known solution, described in the scientific publication by N. Hayakawa, et AL, 2012, entitled "Fabrication Technique of Permittivity Graded Materials (FGM) for Disk-Type Solid Insulator, Proceedings of the CEIDP", is based on the realization of a potential-dimming (or electric-field-dimming) material as an electrically insulating encapsulation material for electronic components. The potential gradation is carried out by a material with a permittivity gradient. This is a composite based on a polymer matrix loaded with particles of different sizes. This composite, before being cured, is subjected to a centrifugal force so as to cause the displacement of the particles in the polymer matrix to obtain a certain permittivity profile as a function of the spatial distribution of the particles.

However, this solution has a certain number of drawbacks. Due to the nature of the technique used to move the particles in the polymer matrix (centrifugation), only unilateral displacement of the particles in the matrix is possible (in the direction opposite to the center of rotation of the material), which is not not optimal. This technique is restrictive because it only makes it possible to treat a single localized area of the material (and not necessarily all of the areas that should be treated (the absence of partial discharges in the material is therefore not guaranteed). ), and further implies that the material has a simple geometric shape (cylindrical or circular). This solution also requires re-machining of the encapsulation material after the centrifugation and hardening steps, which is tedious and expensive to implement. Finally, it does not seem compatible for the manufacture of bulky electrical components, such as power modules, transformers or high voltage switchgear, for example.

Moreover, since the appearance of power modules in the 1990s and despite the rise in power (in terms of current ratings) of this component, there has not been any technological breakthrough in this component by one point view of the voltage rating levels.

Currently, the highest voltage available for silicon-based power modules is 6.5 kV. From an electrical insulation point of view, the main idea to make changing the different voltage ratings was to increase the isolation distance. This approach becomes inoperative after a certain level of tension.

It therefore also appears necessary to propose an innovative electrical insulation solution which makes it possible to effectively reduce the phenomenon of partial discharges in an electrical component and / or to increase the operating voltage, and which is simple and inexpensive to set up. artwork.

Disclosure of the invention

In a particular embodiment of the invention, there is proposed a method of manufacturing a composite part comprising:

- a substrate made of a first material, said first material being a ceramic or a ceramic matrix composite, and

- several three-dimensional volumes disposed in said substrate, each of said three-dimensional volumes being made of a second material which is a ceramic or a ceramic matrix composite having a predefined electrical conductivity, permittivity and thermal conductivity, said three-dimensional volumes being made up of the same ceramic or the same ceramic matrix composite or else different ceramics or ceramic matrix composites,

the method comprising the following steps:

- Making a first preform made of the first material and intended to form the substrate and obtaining several hollow impressions on one face of said first preform, each of said hollow impressions being configured to delimit one of said three-dimensional volumes,

- total or partial filling of each of said hollow impressions with the second material so as to produce second preforms each forming one of said three-dimensional volumes,

- Assembly by sintering of the first preform and of the second preforms so as to obtain the composite part.

The second material can also have predefined mechanical properties. The invention relates to a method of manufacturing a reliable composite part made of a heterogeneous ceramic comprising several (two or more) areas of one. ceramic and / or a ceramic matrix composite having controlled properties and three-dimensional geometry.

It will be recalled that composites with a ceramic matrix are characterized by a set of carbonaceous or ceramic or metallic particles which are incorporated into a ceramic matrix.

The method of manufacturing a part incorporating several volumes comprising the same material or different materials is advantageous in the sense that the steps of the latter are simple to implement, sintering (flash for example) allowing in particular densification and 'simultaneous assembly of two or more different materials, elaboration in a single step and the implementation of a single thermomechanical cycle.

The composite parts manufactured exhibit controlled properties and geometries, good mechanical strength, a high density of up to 99% and a very high degree of reproducibility, up to 100%.

In other words, the proposed manufacturing process is efficient, reliable, simple and fast, but also economical in terms of time and energy.

The technological considerations required for a reliable production of the composite part are in particular:

i) minimization of thermomechanical stresses at the 3D interfaces between the substrate and the volumes which are integrated therein during the manufacture of the composite part. This minimization of stresses involves the development of the material (s) constituting the volumes having a coefficient of thermal expansion similar to that of the substrate.

ii) the chemical stability of the part which implies that the chemical elements of the material (s) of the volumes do not diffuse into the material of the substrate and vice versa. This protects the properties of the different parts of the composite part and consequently ensures the multiplicity of its functions.

iii) optimization of the sintering conditions (temperature, pressure and time) for simultaneous production and assembly of the substrate and the volumes allowing the manufacture of a reliable composite part integrating volumes with controlled properties and geometry. According to a particular implementation, after the filling step and prior to the step of assembling the first preform and the second preforms, the manufacturing process comprises a step of depositing a layer made of the first material so as to cover and embed the second preforms in the first preform.

According to different implementations of the invention, the preforms are obtained by the same technique or different techniques chosen from the techniques of extrusion, pressing of cast strips, deposition by screen printing, deposition by ink jet and additive manufacturing processes.

According to a particular embodiment of the invention, the preforms are obtained by compacting powder.

In a particular embodiment of the invention, the manufacturing process comprises the following steps:

- Placement in a mold of a first layer of the first powder material intended to form the substrate;

- Compaction of the first layer so as to form the first preform and obtaining several hollow impressions on one face of said first preform, each of said hollow impressions being configured to delimit one of said three-dimensional volumes;

- Placement in said mold of a second layer of the second material so as to cover said face of the first preform and fill said hollow indentations to form the three-dimensional volumes;

- Compaction of the second layer so as to form the second preforms;

- Assembly of the first preform and of the second preforms by sintering so as to obtain the composite part;

- Extraction from the mold of said composite part;

- rectification of the second preforms consisting in removing the part of the second preforms situated above said face of the first preform without removing the other part of the second preforms each forming a three-dimensional volume disposed in the first preform.

According to different implementations of the invention, the sintering of the first preform and of the second preforms consists of conventional sintering, such as natural sintering, under load sintering (HP) or isostatic sintering, or in unconventional sintering, such as flash sintering (SPS), laser sintering or microwave sintering. The invention also relates to a composite part obtained by such a manufacturing process comprising:

- several three-dimensional volumes arranged in said substrate, each of said three-dimensional volumes being made of a second material which is a ceramic or a ceramic matrix composite having a predefined electrical conductivity, permittivity and thermal conductivity,

said three-dimensional volumes consisting of the same ceramic or of the same ceramic matrix composite or else of different ceramics or ceramic matrix composites.

According to a particular embodiment of the invention, said three-dimensional volumes are arranged on one face of said substrate.

According to another particular embodiment of the invention, said three-dimensional volumes are arranged within said substrate.

According to one embodiment of the invention, the first material is composed of a mixture of AlN with Xi% m Y ₂ 0 ₃ and x ₂ % m CaF ₂ , the second material being composed of a mixture of AlN with Xi % m Y ₂ 0 ₃ and x ₂ % m CaF ₂ and y vol.% MCG (multilayers of graphene) or of another carbonaceous material, or of a metal disulfide such as MoS ₂ , MXS _2, WS ₂ with 0 , 5 <Xi <6 m%, 1 <x ₂ <7 m% and 0.01 <y <10 vol.%.

According to an alternative embodiment, the first material is composed of a mixture of AlN and sintering additions and the second material is composed of:

of a mixture of AlN and y vol. % of graphene multilayers, or

of a mixture of AlN and y vol. % of carbon nanotubes, or

of a mixture of AlN and y vol. % of carbon,

with 0.01 £ y <10 vol.%.

In another embodiment of the invention, the first material is composed of Al ₂ 0 ₃ and the second material is composed:

of a mixture of AI ₂ 0 ₃ and y vol. % of a metal disulfide and y vol. % of graphene multilayers, or of a mixture of AI ₂ 0 ₃ and y vol. % of a metal disulfide and y vol. % of carbon nanotubes,

with £ 0.01 y £ 10 vol.%.

The metal disulfide can be MoS ₂ , or WoS ₂ or MxS ₂ .

In another embodiment of the invention, the first material is composed of Si ₃ N ₄ or of SiAION and the second material is composed:

of a mixture of Si ₃ N ₄ or SiAION respectively and y vol. % of a metal disulfide and y vol. % of graphene multilayers, or

of a mixture of Si ₃ N ₄ or SiAION respectively and y vol. % of a metal disulfide and y vol. % of carbon nanotubes,

with 0.01 <y <10 vol.%.

The metal disulfide can be MoS ₂ , or WoS ₂ or MxS ₂

The invention further relates to an electrical component comprising such a composite part.

According to a particular application, the electrical component takes the form of an electronic power module, said composite part constituting the insulating substrate of said power electronic module, said three-dimensional volumes arranged in said substrate of said composite part having an electrical conductivity and a permittivity different from those of said substrate so as to reduce the electrical stress in the insulating substrate.

Such an electrical component uses a substrate, in the form of a composite part, with a locally modified conductivity, only at the levels of the integrated volumes.

Such an electrical component can take the form of an electronic power module, said composite part constituting the ceramic substrate insulating said module.

In this case, the method of the invention makes it possible to improve the performance of high voltage power modules for the conversion of electrical energy, in particular in terms of voltage withstand and reliability, by localized action on the electrical conductivity. of the metallized ceramic substrate.

The method makes it possible to reduce the level of the electric field at the level of the triple point by a spreading of the potential lines by intervention in the ceramic substrate (in aluminum nitride (AIN) for example), and in particular by the implementation of several three-dimensional volumes, called pockets, in which the chemical composition and the properties of the ceramic substrate are partially or entirely modified.

These pockets are obtained by developing a compound with the desired electrical properties and by integrating it by volume into a substrate using flash sintering produced by SPS technology (initials of “Spark Plasma Sintering” in English, ie. ie “plasma flash sintering”) preferentially.

These pockets are arranged in areas of the ceramic substrate where unwanted electric field enhancements can appear when the component is energized.

This induces an increase in the dielectric permittivity and / or in the electrical conductivity in a targeted and adapted manner as a function of the desired properties with respect to the electrical component.

Thus, the method of the invention offers a simple and effective solution based on a targeted adjustment of the dielectric permittivity profile and / or of the electrical conductivity of the composite material forming the ceramic substrate of a high voltage power module. The composite material obtained therefore has a dielectric permittivity and / or electrical conductivity profile reducing the electric field reinforcements, sources of partial discharges within the electrical component in operation.

The treatment method according to the invention therefore makes it possible to guarantee better voltage resistance of the electrical component and consequently an increased lifetime or a range of applications extended to higher voltage applications.

Advantageously, the three-dimensional volumes of such an electrical component are obtained by SPS sintering or sintering under load and have an electrical conductivity in the plane perpendicular to the pressing axis greater than the electrical conductivity in the plane parallel to the pressing axis. .

According to a particular implementation, the ratio of electrical conductivities, called anisotropy factor, in the planes parallel and perpendicular to the pressing axis is between 25 and 10 ¹² .

Presentation of figures Other characteristics and advantages of the invention will become apparent on reading the following description, given by way of indicative and non-limiting example, and the appended drawings, in which:

- Figure 1 already described in relation to the prior art, shows an example of an electronic power module known from the state of the art;

- Figure 2 already described in relation to the prior art, shows a partial sectional view of the electrical module illustrated in Figure 1;

FIG. 3, already described in relation to the prior art, presents a known technique for reducing the formation of partial discharges in a power module such as that presented in relation to FIGS. 1 and 2;

- Figures 4 b), c), e), f), h) and i) illustrate various examples of composite parts that can be obtained by the method of the invention for which the three-dimensional volumes are arranged on one face of the substrate;

- Figures 4a), d) and g) illustrate other examples of composite parts having only a three-dimensional volume arranged on one face of the substrate;

FIG. 5 illustrates various examples of composite parts that can be obtained by the method of the invention for which the three-dimensional volumes are arranged on one face of the substrate;

- Figures 6A to 61 illustrate the different steps of a particular solution for manufacturing a composite part according to the invention;

- Figure 7 illustrates an example of temperature (a) and pressure (b) cycles of the sintering step implemented during the manufacture of a composite part according to the invention, and in particular in the solution particular described in relation to Figures 6A to 61;

- Figure 8A is a schematic sectional view illustrating a high voltage power module in which the invention is implemented;

- Figure 8B is a schematic sectional view illustrating another high voltage power module in which the invention is implemented;

- Figure 8C is a schematic sectional view illustrating yet another high voltage power module in which the invention is implemented;

FIG. 9 illustrates the distribution of the electric field and of the equipotentials for a reference structure of a power module; FIG. 10 illustrates the distribution of the electric field and of the equipotentials for a structure of a power module with ceramic substrate integrating a "pocket" zone with controlled electrical conductivity under the triple point, illustrated in FIG. 8B;

FIG. 11 represents the values of the maximum electric fields in the three zones (ceramic, modified ceramic pocket and silicone gel) as a function of the electrical conductivity in the pocket integrated in the ceramic of the structure of FIG. 10;

- Figure 12 shows, for a structure of a ceramic incorporating a pocket (such as Figure 8B), the values of the maximum electric fields as a function of the variation of the conductivity s of the pocket and of the frequency of the sinusoidal voltage d 'food;

- Figures 13 b) and c) illustrate various examples of composite parts according to the invention for which the three-dimensional volumes are arranged within the substrate;

- Figure 13 a) illustrates a composite part for which a single three-dimensional volume is placed within the substrate;

- Figure 14 is a simplified block diagram showing the different steps of the manufacturing process of a composite part according to the invention;

FIG. 15 is a simplified block diagram showing the various steps of the particular manufacturing process of a composite part for which the three-dimensional volume (s) are placed within the substrate;

- Figure 16 shows a composite part to illustrate the value ranges relating to the dimensions of the composite part for the general case and the case of the example of power electronics;

- Figure 17 shows a composite part according to the invention to illustrate the value ranges relating to the dimensions of the composite part for the general case and the case of the example of power electronics;

- Figure 18 illustrates the directions parallel and perpendicular to the pressing axis of a composite part according to the invention obtained by SPS sintering; - Figure 19 illustrates the anisotropy of the electrical conductivity of a composite part according to the invention and obtained by SPS sintering, in the directions parallel and perpendicular to the pressing axis.

5 Detailed description

In all the figures of this document, identical elements and steps are designated by the same numerical reference.

Structure of a composite part and manufacturing process

The inventors have implemented an original process allowing the development of a composite part consisting of a substrate (a functional, homogeneous or heterogeneous monolithic ceramic, or a composite with a ceramic matrix) provided with one (or more) volume (s). ) with controlled properties.

FIG. 4 illustrates various examples of such a composite part 11 which in this case comprises a substrate 111 made of material or compound A integrating a three-dimensional volume 112 (examples referenced a, d and g) or, in accordance with the invention, several three-dimensional volumes 112 (examples referenced b, c, d, e, f, h, i) each made up of a material or compound distinct from material A.

These three-dimensional volumes 112 may be of the same chemical composition (examples referenced b, e, f, h, i) or of different chemical compositions (example referenced c).

FIG. 5 illustrates other examples of such a composite part 11 which comprises a substrate 111 made of material A integrating several three-dimensional volumes 112 each made up of a material B, C, D, E or F distinct from material A.

For all these examples of FIGS. 4 and 5, the materials A to F can be monolithic or composite ceramics.

The respective properties (electrical, thermal, mechanical in particular) of materials A to F and their dimensions are checked.

As can be seen from the examples of Figures 4 and 5, these three-dimensional volumes 112 can take the form of a cylinder or a cube. They can in a variant take the form of a sphere or else take other geometric shapes depending on the envisaged application. It should be noted that different shapes of patterns can be produced including compounds / materials distinct from each other with more complex geometric shapes in surface and in volume than those presented in Figures 4 and 5.

Thus, examples a), b) and d) to i) of FIG. 4 illustrate ceramic parts provided with one or more volumes of composite B with different controlled three-dimensional geometries.

Examples c) of figure 4 and a) to d) of figure 5 illustrate ceramic parts provided with several (two or more) volumes of different composites with a ceramic matrix with a controlled three-dimensional geometry.

FIGS. 13 b) and c) illustrate various examples of composite parts in accordance with the invention for which the three-dimensional volumes are arranged within the substrate.

In the example of FIG. 13 a), a single three-dimensional cylindrical volume 112 made of material B extends inside the cylindrical substrate 111 made of material A.

In the example of FIG. 13 b), several (in this case five in this example) three-dimensional volumes 112 (cylindrical in this example) of material B extend inside the cylindrical substrate 111 of material A.

In the example of figure 13 c), several (in this case five in this example) three-dimensional volumes 112 (cylindrical in this example) of material B, C, D, E and F respectively extend inside of the cylindrical substrate 111 made of material A.

The method developed by the inventors allows three-dimensional integration of these volumes 112 in the substrate 111 with a geometry that is also controlled. The chemical composition of the three-dimensional volumes 112 is adjusted so as to achieve the desired properties on the one hand and to establish reliable three-dimensional (3D) interfaces between the integrated three-dimensional volumes 112 and the substrate 111, on the other hand.

Each three-dimensional volume 112 can be obtained by partially or entirely modifying the chemical composition of the substrate 111.

The production of a composite part 11 therefore requires the development of the material (s) which constitute the three-dimensional volumes 112 and their integration in volume into the substrate 111. The choice of the constituent elements of this material (B , C, ...) is not only decisive in providing the composite part 11 with the functions technological requirements, but it is also subject to several technological criteria decisive for a reliable integration of this material.

Indeed, in addition to its functional properties which differ from those of the substrate 111, the material or compound of an integrated three-dimensional volume 112 must:

have sintering conditions similar (atmosphere, temperature ranges, pressure ranges) to those of compound or material A of substrate 111 in order to be able to perform simultaneous consolidation of materials A, B, C, etc. and sinter welding from these to each other at three-dimensional interfaces;

be made up of stable chemical elements or compounds which do not diffuse or in a controlled manner through the three-dimensional interfaces of one of the materials towards the other, in particular at the manufacturing temperature (flash sintering called "SPS") of the composite part 11 and that of its use. In other words, the substrate 111 must act as a diffusion barrier for the constituent elements of the integrated materials or compounds and vice versa. This makes it possible to avoid or at the very least limit the progressive degradation of the functional properties of the materials (A, B, etc.) and consequently to preserve the functioning or the technological functions of the composite part 11;

have a coefficient of thermal expansion (CDT) similar to that of substrate 111. This minimizes, at the interfaces of materials (A / B, A / C, etc.), the thermomechanical stresses induced by the thermal cycles to which the composite part 11 is exposed during manufacturing by flash sintering (SPS) and during its use, and thus prevents its cracking, in particular at three-dimensional interfaces, which would lead to irreversible damage to the composite part 11.

The manufacture of a composite part 11 also requires the prior determination of the shrinkage after sintering of each of the materials, in particular of the material or compound B (composite with ceramic matrix) integrated in the material A (ceramic), in order to control the dimensions. and the geometry of the pattern to be integrated into the substrate 111 of material A. The main steps of the manufacturing process of a composite part, such as those illustrated in FIGS. 4, 5 and 13. This composite part can comprise one or more three-dimensional volumes at the sides of FIG. controlled electrical and / or thermal and / or mechanical properties.

In this example, a composite part 11 is produced consisting of a heterogeneous substrate 111 made of material A (AIN + Xi% m Y ₂ 0 ₃ + x ₂ % m CaF ₂ ) incorporating a three-dimensional volume 112 made of material B (AIN + Xi% m Y ₂ 0 ₃ + x ₂ % m CaF ₂ + y vol.% MCG), with 0.5 £ Xi £ 6 m%, 1 <x ₂ <7 m% and 0.01 £ y £ 10 vol.% (in percentage of mass m% or volume vol.%).

The respective fractions of yttrium oxide Y ₂ 0 ₃ (xi) and calcium fluoride CaF ₂ (x ₂ ) were determined based on the values reported in the literature, but also on an experimental optimization to achieve a maximum densification combined with the highest possible thermal conductivity of the AlN-based ceramic (material B). A range of fraction values is given here for each of the additions: 0.5 - 6 m% for Y ₂ 0 ₃ and 1 - 7 m% CaF ₂ . Note that these fractions depend, on the one hand, on the quality of the starting AIN powder, in particular with regard to the level of oxygen (impurities) contained in it, and on the other hand, on the experimental conditions of the sintering step detailed below.

In a first step (S30), a first PI preform is made consisting of the first material A and intended to form the substrate 111, and at least one hollow imprint is obtained on one face of said first PI preform, each hollow imprint being configured to delimit a three-dimensional volume 112.

Said at least one hollow cavity is then completely or partially filled (S31) with the second material B so as to produce at least one second preform P2 forming at least one three-dimensional volume 112, then the first preform PI is assembled (S33) and the second preform or second P2 by sintering so as to obtain the composite part 11. After the assembly step (S33), a step of leveling off a surplus of material B can be implemented.

The main steps of a particular process, by powder compacting, for manufacturing a composite part, such as those illustrated in Figures 4 and 5, are described below with reference to FIGS. 6A to 61. In this particular process, a powder of a first material A is loaded into a graphite mold 30 of diameter equal to 20 mm as illustrated in FIG. 6A. The mold 30 has an internal shape corresponding to the shape of the part to be obtained (cylindrical in the present case). A graphite sheet was previously placed on the internal wall of the die and above the graphite piston 31 in order to avoid any reaction between the enclosure and the first material A.

Then, is introduced into the mold 30 a piston 32 with a stainless steel imprint of diameter slightly less than 20 mm which is provided at its end directed towards the mold with a cylindrical pattern, or relief, 321 of diameter equal to 14 mm and of thickness "e" equal to 600 μm (FIG. 6B).

In the next step, a uniaxial cold pressure (FIG. 6C) is applied to the cavity piston 32 so as to compact the powder of material A and etch a hollow cylindrical EM imprint (with a diameter equal to 14 mm and d thickness "e" equal to 600 μm) in the preform PI (FIG. 6D). The shape of the hollow EM imprint is given by that of the piston 32 and defines a three-dimensional volume of the composite to be obtained.

The preform P2 is then made, made of a second material B, above the preform PI. Note that the electrical conductivity o _A of material A is lower than the electrical conductivity o _B of material B.

After removing the stainless steel piston 32, the powder of material B is poured onto the preform PI in the mold 30 so as to fill the hollow EM cavity and cover the entire upper face of the preform PI (FIG. 6E).

A second graphite piston 33 is then introduced above the powder of material B to close the mold 30 and a uni-axial cold pressure is applied to this piston 33 (FIG. 6F). This makes it possible to compact the powder of material B and to obtain the preform P2.

The flash co-sintering (“SPS” for “Spark Plasma Sintering”) of the preforms PI and P2 (FIG. 6G) is then carried out in a controlled atmosphere at a temperature T _Sp s of between 1550 and 1850 ° C. for a period t _SP s preferably between 1 and 20 minutes under a pressure P _SP s between 30 to 100 MPa with a rise and / or fall in temperature vc between 30 to 200 ° C / min. During this treatment, the preforms PI and P2 are under pressure from the upper piston 33 and the lower piston 31. Particular temperature and pressure cycles (a) and (b) of this sintering step are illustrated in Figure 7.

It is recalled here that the SPS technology combines, simultaneously, the application of a high uniaxial pressure and high intensity direct current pulses causing an almost immediate and uniform temperature rise. Besides the time and rate of heating, the most important parameter for the SPS sintering process is the sintering temperature. The two preforms to be assembled may be of the same ceramic or else each of the parts to be assembled may be of a different ceramic (which is the case here). The surfaces to be assembled are then brought into contact without any addition of any kind being placed between these surfaces. Pressure contact is established between the two facing surfaces to be assembled. Then, while maintaining the pressure, a pulsed electric current is applied to said preforms so as to raise the temperature of the preforms. In other words, when everything is in contact, an electric current is generated in order to create the temperature rise. Once the assembly is complete, the application of the electric current and the pressure is stopped, and the preforms are cooled.

The flash sintering treatment allows not only the co-sintering of the two preforms PI and P2, but also the welding of the latter to obtain a single dense composite part (FIG. 6H).

In other words, at the end of the treatment by flash sintering, a dense composite part is thus obtained for which the second composite material B is integrated into the first material A by welding obtained by sintering.

The assembled preforms are extracted and the depth of the pattern after sintering and the amount of material B to be removed are then evaluated.

The second preform P2 is then rectified, preferably by polishing, consisting in removing the part of the second preform P2 located above said face of the first preform PI without removing the other part of the second preform forming a three-dimensional volume disposed in the first preform.

After polishing, a composite part 11 of good mechanical strength is obtained (FIG. 61) having a substrate 111 provided with a cylindrical volume 112 at its center, this volume consisting of the composite with controlled electrical and thermal conductivities. It will be understood that these steps are also implemented when several three-dimensional volumes are placed in the first preform.

This composite part 11 produced by co-sintering the two preforms PI and P2 is obtained by applying a single thermomechanical cycle, which represents a significant saving in terms of time and energy. Observation with an optical microscope revealed the absence of cracks at the interfaces of materials A and B (the TDCs of ceramics are similar) suggesting the presence of low thermomechanical stresses, much lower than the breaking stress of materials A and B, nor diffusion of material B to material A.

Examination of a polished cross section of this composite part 11 made it possible to measure the depth of material B after sintering. It is of the order of 150 μm, which indicates a substantial shrinkage, ie 450 μm, of the entire part.

As regards the composite parts 11 of FIG. 13 for which one or more three-dimensional volumes 112 are embedded within the substrate 111, the manufacturing process comprises the following steps illustrated in FIG. 15. In a first step (S30), a first PI preform is made consisting of the first material A and intended to form the substrate 111, and at least one hollow imprint is obtained on one face of said first PI preform, each hollow imprint being configured to define a three-dimensional volume 112.

Said at least one hollow cavity is then completely or partially filled (S31) with the second material B so as to produce at least one second preform P2 forming said at least one three-dimensional volume 112.

The method then comprises a step (S32) of depositing a layer of first material A so as to cover and embed said at least one second preform P2 in the first preform PI, then the first preform PI and said preform are assembled (S33). at least a second preform P2 by sintering so as to obtain the composite part 11.

Whatever the manufacturing process envisaged to obtain a composite part comprising one or more three-dimensional volumes located on one face of the substrate or else embedded in the latter, the flash co-sintering of the preforms PI and P2 is carried out in a controlled atmosphere at a temperature T _SP s between 1550 and 1850 ° C for a period t _SP s preferably between 1 and 20 minutes under a pressure P _SP s between 30 to 100 MPa with a rise and / or fall in temperature vc between 30 to 200 ° C / min, for a PI preform made of material A (AIN + x ₄ % m Y ₂ 0 ₃ + x ₂ % m CaF ₂ ) and a P2 preform made of material B (AIN + x ₄ % m Y ₂ 0 ₃ + x ₂ % m CaF ₂ + y vol.% MCG with 0.5 £ Xi £ 6 m%, 1 £ x ₂ £ 7 m% and £ y £ 10 vol.% (As a percentage of mass m% or volume vol%).

It should be noted that the sintering of the PI and P2 preforms can be obtained, in a variant, by conventional sintering, such as natural sintering or under load (called "HP" for "Hotpressing" in English), or isostatic sintering, or unconventional sintering, such as flash sintering (SPS), or laser sintering, or microwave sintering.

It should be noted that the methods for producing the preforms PI and P2 may be different from that described above, consisting of powder compacting. Thus, the PI and P2 preforms can be obtained by the same technique or by different techniques (hybrids) chosen from the techniques of extrusion, pressing of cast strips (polymer / ceramic composite), deposition by screen printing, deposition by jet. ink and any additive manufacturing technique.

Description of the process in the context of an application to power electronics An example of interest of this process, detailed below, is the production of a composite part based on aluminum nitride (AIN) for power electronics applications.

The invention is of course not limited to this example (geometries and / or properties), nor to this particular field of application of power electronics.

This method makes it possible to improve the performance of high voltage power modules for the conversion of electrical energy in terms of voltage withstand by localized action on the electrical conductivity and permittivity of the metallized ceramic substrate.

Conventionally, a power module consists of two main insulating materials: the ceramic substrate (Type Al ₂ 0 ₃ , Si ₃ N ₄ , AIN) and the encapsulating medium (most often a silicone gel or an epoxy resin). As pointed out before, the interface of these two insulators is one of the weak points of the power module, and more precisely the junction between the substrate, the metallization and the encapsulation which is a zone of increase of the electric field. This reinforcement zone constitutes a critical point of the insulation system where the electric field is the highest.

The invention makes it possible to act by reducing the level of the electric field at the triple point by a spreading of the potential lines. This spreading is obtained by acting on the electrical conductivity by incorporating a conductive additive in the ceramic material forming the substrate.

In other words, the invention consists in locally modifying the properties of the ceramic material at the level of the triple point by controlling the electrical conductivity and permittivity in several predefined three-dimensional geometric zones.

Three implementations are proposed in Figure 8A, 8B and 8C, the first using a striped modified conductivity layer on the substrate (Figure 8A), the second in which a modified electrical conductivity pocket is located under each triple point (Figure 8B), and the third in which said modified ceramic pocket is integrated under each triple point and extends under the metallization and not only at the edges (FIG. 8C). Each of these pockets corresponds to a three-dimensional volume obtained according to the method described above.

For the second and third implementations of FIGS. 8B and 8C, the first ceramic material A is, for example, based on aluminum nitride (AIN).

It is thus proposed to add to the ceramic powder, here AIN, a conductive additive (called a conductive addition) making it possible to locally modify the electrical properties of the material so as to provide one or more pockets of ceramic material B - here modified AIN ( electrically) -, but also to propose an adapted object geometry and to shape the modified AIN / AIN assembly using a sintering method (flash for example) described in detail previously.

Thus, the process for producing composite parts makes it possible to manufacture a heterogeneous AIN ceramic substrate integrating in volume one or more pockets with controlled electrical and thermal properties.

The approach of the invention consists in acting on the properties of the ceramic and makes it possible to give the ceramic substrate new properties at the heart of the latter.

Material B includes AIN aluminum nitride which is modified by incorporating: - a sintering additive to promote sintering in the liquid phase of the material during the SPS treatment and to remove the oxygen impurities in the material: it can be yttrium oxide Y ₂ 0 ₃

- an additive to reduce the quantity of secondary phases formed at the grain boundaries during sintering: it may be calcium fluoride CaF ₂

- conductors such as multilayers of graphene (MCG) in order to take advantage of the mobility of high charge carriers of the graphene sheets: this makes it possible to adjust the electrical properties and in particular to play on the electrical conductivity and permittivity of material B (Modified AIN).

According to a particular embodiment, the substrate can comprise on one of its faces several volumes, called here pockets, with controlled electrical and thermal properties. These pockets are formed on one face of the substrate by the manufacturing method described above, by the implementation of a particular relief on the impression piston when the solution described in relation to FIGS. 6A to 61 is implemented.

According to another particular embodiment, the interior of the substrate can comprise several three-dimensional volumes, or pockets, with controlled electrical and thermal properties. In other words, these three-dimensional volumes are buried within the substrate.

Observations carried out by the inventors show that the compound / material B displays, for the relatively high fractions of MCG (for example at 2.5 vol%), an electrical conductivity insensitive to the frequency, indicating a resistive behavior over the entire range of frequency explored. The samples produced display almost identical values revealing a remarkably high degree of reproducibility. This increase in electrical conductivity is linked to the presence of a high number of graphene monolayers endowed with a higher mobility of charge carriers, which induces more interesting electrical effects in the ceramic matrix AIN.

Compound / material B displays, for smaller fractions of MCG (eg at 1.25 vol%), a frequency dependent electrical conductivity, indicating capacitive behavior over the entire frequency range explored. The incorporation of these compounds in the ceramic based on AIN by co-sintering causes a local increase (in the pockets) of the electrical conductivity of the latter. Thus, the object of the invention is to modify the substrate and in particular to locally modify (at the triple point "ceramic (based on AIN) insulating / metallization / silicone gel") the electrical conductivity so as to reduce the field at triple point.

This approach is implemented before the metallization of the ceramic whereas, in the prior art, it is proposed to act by modifying the surface after the metallization of the ceramic.

In addition to the technological functions (electrical insulation, thermal conduction (heat extraction), support and mechanical strength) provided by a conventional substrate in a power module, the composite part described above can fulfill the function of the gradation of the electric or potential field. in such a module by controlling the electrical conductivity and permittivity in volumes integrated in the substrate. The approach adopted for this process makes it possible to produce a completely dense part, exhibiting excellent mechanical strength.

The use of a ceramic substrate at the present time remains essential in high voltage power modules, in particular from the point of view of heat transfer. Despite everything, the thermal dimensioning strongly limits the increase in the thickness required to withstand an increase in the level of voltage withstand. The solution presented makes it possible to optimize the compromise existing between “high voltage withstand” and “high thermal dissipation” of the structure under application conditions, a compromise which is based on the thickness of the ceramic.

By way of illustration, a first example of the structure of a composite substrate for a power electronics module is given for which a cylindrical pattern made of a composite material (material B) with controlled electrical conductivity and permittivity is integrated into the heterogeneous insulating ceramic (material A).

In this first example, ceramic A has a diameter of 20 mm and a thickness of 2 mm. It is composed of AIN + Xi% mY ₂ 0 ₃ + x ₂ % m CaF ₂ . Composite B has a diameter of 14 mm and a thickness of 150 µm. It is composed of AIN + Xi% m Y ₂ 0 ₃ + x ₂ % m CaF ₂ + y vol.% MCG (multilayers of graphene) or y vol.% Of another carbonaceous material, of MoS ₂ , or of WS ₂ . We give here a range of fraction values for each of the additions: 0.5 - 6 m% for Y ₂ 0 ₃ and 1 - 7 m% CaF ₂ and 0.01 £ y £ 10 vol.% (in percentage of mass m% or volume vol.%) .

By way of illustration, other examples of the structure of a composite substrate for a power electronics module are given for which a cylindrical pattern in a composite with a ceramic matrix (material B) with controlled electrical conductivity and permittivity is given. integrated in the insulating ceramic (material A).

For all these examples, the electrical conductivity o _A of ceramic A is lower than the electrical conductivity o _B of ceramic B.

According to a particular implementation, ceramic A is composed of AIN and sintering additions (rare earths and / or alkaline-earth oxides, such as Y ₂ 0 ₃ , CaF ₂ ). In this case, the composite B can be composed of AIN + y vol. % MCG (multilayers of graphene), or of AIN + y vol. % CNT (carbon nanotubes) or AlN + y vol. % carbon with 0.01 £ y £ 10 vol.% (in percentage of volume vol.%).

According to another particular implementation, ceramic A is composed of Al ₂ 0 ₃ . In this case, the composite B can be composed of Al ₂ 0 ₃ + y vol. % MCG (multilayers of graphene), or of Al ₂ 0 ₃ + y vol. % CNTs (carbon nanotubes) with 0.01 <y <10 vol.% (In percentage of volume vol.%).

According to yet another particular implementation, ceramic A is composed of Si ₃ N ₄ or SiAION. In this case, the composite ceramic B can be composed of Si ₃ N ₄ or SiAION respectively + y vol. % MCG (multilayers of graphene), or of Si ₃ N ₄ or SiAION respectively + y vol. % CNTs (carbon nanotubes) with 0.01 <y <10 vol.% (In percentage of volume vol.%).

The “y” fraction corresponds to that of the conductive addition (MCG, CNT, carbon, MoS ₂ or WS ₂ ) which is introduced into the compound (AIN, Al ₂ 0 ₃ , Si ₃ N ₄ or SiAION) to modify its electrical conductivity and permittivity. The “y” fraction is optimized to achieve the desired electrical conductivity in the composite B which is integrated in the pockets of the ceramic A with the aim of reducing the electric field at the triple points.

Multilayers of graphene (MCG) can be obtained by exfoliation of graphite in isopropyl alcohol, preferably by means of the sonotrode.

The experimental conditions of the exfoliation are selected and the duration of this operation is between 1 and 12 h so as to produce MCGs with a quality allowing, on the one hand, to reach the value of the electrical conductivity at a low level of MCG and, on the other hand, to promote the most uniform possible distribution in the ceramic matrix AIN, necessary for the reproducibility of the results. A low level of MCG will also help to preserve thermal conductivity.

The choice of isopropyl alcohol as solvent for the exfoliation of graphite is governed by its low evaporation temperature (82.6 ° C.) allowing the solvent to be easily separated from the final mixture of the powder.

The powder (AIN + sintering additions) is prepared from the mixture of AIN + Xi% m Y ₂ 0 ₃ + x ₂ % m CaF ₂ which is homogenized in absolute alcohol (dispersing solvent) using a mechanical mixer adjusted to a rotation preferably between 100 and 500 rpm for a period of between 1 and 5 hours. The powder is then heated at 80 ° C for 15 h to remove the solvent.

The powder (AIN + Xi% m Y ₂ 0 ₃ + x ₂ % m CaF ₂ ) + y vol.% MCG is prepared as follows: the graphene multilayers (MCG) dispersed in isopropyl alcohol are introduced with the AIN mixture + Xi% m Y ₂ 0 ₃ + x ₂ % m CaF ₂ in a glass flask which was then fixed in the rotary evaporator. The flask is subjected to a rotation of between 50 and 100 rpm for a period of between 1 and 5 hours to homogenize the solution. The system is then placed under vacuum to reduce the partial pressure of the solvent. The vapor produced is sucked in, then condensed using a cooling circuit, and the liquid solvent is finally conveyed to a recovery tank.

Heterogeneous ceramics and nanocomposites with a ceramic matrix are produced as follows: the powder mixtures prepared (AIN + Xi% m Y ₂ 0 ₃ + _x2 % m CaF ₂ ) and (AIN + Xi% m Y ₂ 0 ₃ + x ₂ % m CaF ₂ + y vol.% MCG) are treated by flash sintering (SPS) in a controlled atmosphere with the forming conditions: T _SP s = 1550 - 1850 ° C, P _SP s = 30 - 100 MPa, t _SP s = 1 - 20 min and vc = 30 - 200 ° C / min.

Simulations allowed a quantitative measurement of the effect of the proposed solution on the distribution of the electric field in materials in the vicinity of the triple point in a power electronics module.

As a reference, the results obtained in the case of a 4.8 kV voltage rating power module, made up of an AIN substrate (dielectric permittivity: E _A I _N = 8.8; electrical conductivity: s _{A | N} = ÎO ¹³ S / m), 635 miti thick, encapsulated by a silicone gel (e _gei = 2.7; o _gei = 10 ¹³ S / m). The copper electrode on the upper face (300 μm thick) is connected to the source delivering an alternating voltage of 4.8 kV at 50 Hz (reference frequency with respect to the results of the bibliography, it is also that of the partial discharge (PD) threshold test according to IEC 1287). The electrode on the lower face is connected to ground. The geometry of the edge of the copper electrode was simulated on the basis of an actual etching profile (measured from a sectional view of a commercial substrate wafer): it appears that this profile is a very important factor. influence the quantitative result of the field.

FIG. 9 illustrates the distribution of the electric field and of the equipotentials for a reference structure with a homogeneous ceramic substrate. We visualize a concentration of the equipotentials around the triple point (electrode / silicone gel / ceramic AIN) and a maximum electric field of 55.5 kV / mm, located in the ceramic, to which the arrow in FIG. 9 points. The electric field is represented by a gradient of colors, the highest values correspond to the warmest colors and the weakest fields to the cooler colors. The equipotential lines are those represented by lines. The value of the maximum field in the gel near the triple point is evaluated at 50.7 kV / mm.

FIG. 10 illustrates the distribution of the electric field and the equipotentials for a structure with a ceramic substrate with controlled conductivity in a zone "with pocket" of modified ceramic and makes it possible to visualize the effect of a pocket with controlled electrical conductivity (e _{ro <} ± _q = 8.8; o _poChe = _1.8 × 10 ⁷ S / m), depth 50 μm and width 500 μm, on the distribution of equipotentials and the values of the electric field in the vicinity of the triple point.

For this particular value of electrical conductivity of the ceramic in the pocket, it was observed:

- a decrease in the maximum electric field at the triple point (PT Zone, the 'triple point' in figure 10) under the electrode of nearly 50% (28 kV / mm against 55 kV / mm for the homogeneous substrate) ,

- a "sharing" of the maximum stress in the field with other zones, in particular at the edge of the pocket (zone R of 'transfer' in FIG. 10), where a reinforcement of the field is obtained on the surface in the gel (with a maximum value of 26 kV / mm) and in depth in the ceramic (with a maximum value of 28 kV / mm).

This reduction in the maxima of the electric field is of major interest in terms of improving the reliability and increasing the voltage of electronic power devices.

The distribution of the electric field is dependent on the electrical conductivity of the modified ceramic pocket and this dependence is a function of the frequency of the voltage applied.

The influence of the o _poChe parameter is given by the simulation results presented in figure 11 where the values of the maximum electric fields are reported in the three zones (ceramic (AIN), pocket (modified AIN) and silicone gel) as a function of the electrical conductivity in the modified ceramic pocket (on the abscissa) of the structure of figure 10.

It is observed for a sinusoidal excitation at 50 Hz of amplitude 4.8 kV that:

- the pocket becomes influential for a conductivity greater than lxlO ⁸ S / m, value from which the maximum field in the pocket and in the gel begins to decrease in the PT zone (figure 10) of the triple point, and the maximum field in the ceramic and the gel in the zone R (figure 10) at the edge of the pocket begins to increase,

- for a particular value of the conductivity close to 2xl0 ⁷ S / m, a balance is obtained between the values of the field peaks in the zones PT and R: this is the case presented previously in figure 10 for which the maximum of field in the whole insulating structure is the weakest (28 kV / mm),

- for a conductivity in the pocket higher than this value, the reinforcement of the field is transferred entirely to the zone R (figure 10), with values of the maximum fields in the ceramic (41 kV / mm) and in the gel (33 kV / mm) which increase compared to the optimal case, but which are in all cases reduced compared to the case without pocket (corresponding here to the case with o _pocket = 10 ¹³ S / m). It should be noted, however, that the intensity of the field reinforcement in this zone R will also depend on the profile of the edge of the pocket.

The simulation thus makes it possible to know the range of electrical conductivity favorable to the reduction of the field reinforcement in the structure for operation at a given frequency. It also shows that this range is extended, o _poChe to be greater than a certain minimum value (2xl0 ⁷ S / m, for 50 Hz), for which all the maximum fields of the structure will have been reduced, and beyond which, we will benefit at least from the suppression of the field peak at the metallization edge (point PT), or even also of a maximum field peak in the gel compared to the case with a homogeneous substrate (FIG. 9), which is smaller and at the edge of the pocket.

In summary, this digital study shows that the range of electrical conductivity to be obtained, in order to be able to benefit from a significant reduction in the electrical stress on the insulators of the module, is wide. It also demonstrates a dual interest of the proposed concept, which allows:

- a transfer of the field stress to the end of the pocket, far from the area for improvement in the vicinity of the triple point (electrode etching defects, irregular etched metal edges, solder burrs, cavities in the encapsulant);

- a very significant gain in the maximum field stress undergone by the insulators: ie 50% for optimum electrical conductivity in the pocket, and up to 100% in the area of the triple point PT; ie 25% for the ceramic, and 35% for the gel in the transfer zone R, for any conductivity in the pocket greater than this optimum.

In the context of a structure with a pocket with controlled electrical conductivity (FIG. 10), a variation in the frequency of the sinusoidal supply voltage was carried out as a function of the variation in the conductivity of the pocket. The result is shown in figure 12.

These results indicate that, for a given structure (given dimensions and material properties), the properties s, e 'and e "of the modified AIN material of the pocket must be chosen according to the maximum operating frequency in the application. aimed, in order to ensure effective protection (ie reduction of the peak field at the triple point) whatever the conditions of use.

The steps for making a power module with IGBT, MOSFET ("Metal Oxide Semiconductor Field Effect Transistor"), or JFET ("Junction Field Effect Transistor") transistors, such as the one partially illustrated in FIG. 8B, are highlighted. work according to well-known techniques from microelectronics.

Its structure consists of a stack of different elements.

A first electrical contact, high potential, connected to a high voltage supply, and a second electrical contact, low potential, connected to ground are arranged on the ceramic substrate on the surface of which have been formed pockets of ceramic material with controlled conductivity and permittivity, these pockets being designed so as to reduce the formation of partial discharges when the component is under voltage. Another electrical contact (not shown in the figure) is arranged on the underside of the layer of ceramic material and connected to ground. These elements form the metallized ceramic substrate. The module further comprises a support made of copper (Cu) or of an aluminum-silicon carbide alloy (AlSiC), commonly called a sole on which the metallized ceramic substrate is placed.

These elements are covered with a layer of electrically insulating protective (or encapsulating) material.

The method according to the invention therefore makes it possible to guarantee the manufacture of an electrical component having better resistance to electrical voltage, fewer partial discharges and consequently an increased lifetime.

FIGS. 16 and 17 show two types of composite parts making it possible to illustrate the ranges of values relating to the dimensions of the composite part for the general case and the case of the example of power electronics.

For the general case, the diameter d _A of the substrate (here of cylindrical shape) of the composite part is between 8 and 80 mm, the height h of the substrate is between 0.5 and 50 mm. By way of example, the diameter d _A of the substrate is equal to 20 mm and the diameter d _B of the three-dimensional volume is 14 mm. The minimum diameter d _B of the three-dimensional volume is between 1 and 2 mm and the maximum diameter d _B of the three-dimensional volume is between the minimum d _B value and 18 mm. The depth e of the three-dimensional volume is between 5 μm and the value of the height h of the substrate. The distance d _m between the three-dimensional volumes is greater than or equal to a value between 1 and 2 mm.

For the case of a composite substrate of a power module, the diameter d _A of the substrate (here of cylindrical shape) of the composite part is 130mm x 180mm (Master Card - Substrate), the height h of the substrate is included between 0.127 and 2 mm, the diameter d _B of the three-dimensional volume ranges from 4mm x 4mm to 125mm x 175mm. The depth e of the three-dimensional volume is less than half of the height h of the substrate. The distance d _m between the three-dimensional volumes is greater than or equal to 2 mm and depends on the level of tension. The process described above is intended for the manufacture of electrical power modules and components. It is clear, however, that it can easily be adapted to other applications.

The applications of the solution developed by the inventors are diverse and relate, for example, to ceramic substrates for power electronics in application to the conversion of high and very high voltage energy (beyond 500V) in the fields of transport (automobile, train, tramway, metro, naval and avionics) and electricity production (wind and photovoltaic development and the emergence of continuous HVDC electrical networks).

Other ceramic materials (Al ₂ 0 ₃ , Si ₃ N ₄ , for example) having different physical properties can be used for the application in power electronics and local optimization in volume of other types of properties (thermal, mechanical, optical, etc.) using other carbon, ceramic or metallic particles.

The modification on a specific imprint of the electrical properties of ceramic could, moreover, be applied to other fields of electrical engineering, such as for example connectors and cable passages.

The process can be used to produce conductive ceramic tracks instead of applying metal tracks (copper, silver, etc.) to insulating substrates. In this case, the conductive ceramic tracks can be integrated in the vicinity of the surface of the substrate, which allows use at high temperature which is much sought after in electronics.

Besides the electronic field, these ceramic contacts can be used in the field of the conversion of thermal energy into electricity by means of thermoelectric generators (TE). Indeed, the process will make it possible to apply these contacts to the ends of semiconductor ceramics produced in the form of TE legs (bars). This would help reduce the electrical and thermal resistance at the interfaces (often very high at the ceramic / metal leg interfaces) and consequently preserve the efficiency of the generator. Ceramic contacts would also enhance the reliability of TE legs at high temperatures. In the same field of thermoelectric generators, the process will also make it possible to manufacture ceramic TE legs fully integrated in a protective ceramic. This will serve to protect many TE materials which are unstable in air and at operating temperature. Many materials are concerned (example: Skutterudite materials).

The integration of passive components requires materials whose functions can be very varied (conductors, insulators, dielectrics and magnetic). Today, most technologies are based on the assembly of discrete components, but in order to gain in power density, functions are increasingly shared in a reduced volume. The ultimate goal is to obtain a single element which integrates most, if not all of the functions of the liabilities. The proposed process opens the way to ceramic structures integrating the four functions on the same object which will require 3D structures, initially, to integrate the connections (conductive materials) within the insulators or dielectrics, and which will ultimately allow a new geometric design of magnetic and / or dielectric functions.

The advantages of the process are as follows:

three-dimensional 3D integration of one or more compounds,

mastery of integrated patterns,

simple and easy to implement steps,

simultaneous densification and welding by SPS sintering of two or more materials of different compositions,

processing in a single step in the case of SPS flash sintering,

possibility of sintering using other shaping techniques (e.g. conventional hot pressing (Hotpressing - HP)),

application of a single thermomechanical cycle, which represents a significant saving in terms of time and energy.

The advantages of the developed compound are as follows:

controlled electrical conductivity,

preserved thermal conductivity,

controllable mechanical properties,

stable to the sintering conditions of the substrate.

The advantages of manufactured composite parts are as follows:

parts with multiple and varied functions, controlled properties and geometries,

mastered 3D interfaces,

very high degree of reproducibility, up to 100%,

very high density, up to 99%,

good mechanical strength.

For power electronics applications, the use of a composite part can lead to a reduction in constraints (in particular electrical, thermal, thermo-mechanical), such as for example a reduction in the value of the maximum electric field by 50% .

Anisotropy of electrical conductivity

The inventors have observed that there is a direct link between the exfoliation conditions (quality of the particles, number of sheets, size of sheets, concentration) and the anisotropy factor (the anisotropy factor being the ratio of conductivities in the planes parallel and perpendicular to pressing), this factor being more particularly improved for composites sintered uniaxially by SPS sintering. It should be remembered that SPS technology combines, simultaneously, the application of a high uniaxial pressure (figure 18) and high intensity direct current pulses causing an almost immediate and uniform rise in temperature.

The results of the electrical measurements show a very large difference between the electrical conductivities in the direction parallel and perpendicular to the pressing. The values of the anisotropy factor, measured at 50 Hz, can reach a factor between 150 and 10 ¹² , particularly between 10 ³ and 10 ¹⁰ , more particularly between 10 ³ and 10 ⁸ and more particularly still between 10 ³ and 10 ^e .

The results show that it is a fine control of the conditions of exfoliation and of the concentration of particles (MLG (Multi Layer Graphene), BN, MoS2, W, among others) in the ceramic which allows these factors to be achieved. 'anisotropy.

In the example of the intended application (ceramic substrate), this property is very relevant because it allows the potential to be graded, only in the useful direction and not to the detriment of the insulation in the thickness of the ceramic.

The anisotropy factor can therefore be modulated through the exfoliation conditions. The anisotropy can be adapted according to the production process, the factor included between 10 ³ and 10 ^s (measured at 50 Hz) being obtained by sintering under load or flash sintering in particular.

The very important anisotropy factors that the inventors have measured make it possible to have a material with conductive behavior in the direction s perpendicular to the pressing, and a dielectric or insulating behavior in the direction sm parallel to the pressing as illustrated in FIG. 18.

Figure 19 shows the electrical conductivity parallel and perpendicular to the pressing axis for a sample AIN + 1 m% Y203 + 2 m% CaF2 + 2.5 vol% MCG, with a total exfoliation time of 2h + 2h + lh. As illustrated in this figure, the anisotropy factor is here equal to 2.63 x 10 ³ .

This opens the way to an additional degree of freedom compared to the 3D geometry of the patterns. For example, the potential gradation could preferably take place in the metallization plane (in a metallized substrate) which makes it possible to keep the dielectric or insulating property in the thickness of the substrate. This approach allows the optimization of the development of a ceramic matrix nanocomposite including multilayers of graphene (Ex. AiN - MCG), integrated into the volumes of the part, displaying an anisotropy of electrical conductivity never equaled to date. This is achieved through careful optimization of the production of graphene multilayers by exfoliation of graphite. Such an anisotropy is controllable and gives a conductivity with two main components, a low conductivity, close to that of the substrate, in a first direction and a high conductivity in a second direction perpendicular to the first direction.

Claims

CLAIMS 1. A method of manufacturing a composite part (11) comprising:

a substrate (111) made of a first material (A), said first material (A) being a ceramic or a ceramic matrix composite, and

a plurality of three-dimensional volumes (112) disposed in said substrate (111), each of said three-dimensional volumes (112) being made of a second material (B) which is a ceramic or a ceramic matrix composite having electrical conductivity, permittivity and predefined thermal conductivity, said three-dimensional volumes (112) being made of the same ceramic or the same ceramic matrix composite or else of different ceramics or ceramic matrix composites,

the method comprising the following steps:

production (S30) of a first preform (PI) made of the first material (A) and intended to form the substrate (111) and obtaining several hollow impressions (EM) on one face of said first preform (PI), each of said hollow impressions (EM) being configured to delimit one of said three-dimensional volumes (112), total or partial filling (S31) of each of said hollow impressions (EM) with the second material (B) so as to produce second preforms (P2) forming each one of said three-dimensional volumes (112),

assembly (S33) by sintering of the first preform (PI) and of the second preforms (P2) so as to obtain the composite part (11).

2. Manufacturing process according to claim 1, characterized in that, subsequent to the filling step (S31) and prior to the assembly step (S33) of the first preform (PI) and of the second preforms (P2), the method comprises a step (S32) of depositing a layer consisting of the first material (A) so as to cover and embed the second preforms (P2) in the first preform (PI).

3. Manufacturing process according to claim 1 or 2, characterized in that the preforms (PI, P2) are obtained by the same technique or techniques various techniques chosen from extrusion, cast strip pressing, screen printing deposition, inkjet deposition and additive manufacturing processes.

4. Manufacturing process according to claim 1 or 2, characterized in that the preforms (PI, P2) are obtained by compacting powder.

5. The manufacturing method according to claim 4, characterized in that it comprises the following steps:

- Placement in a mold (30) of a first layer of the first powder material (A) intended to form the substrate (111);

- compacting the first layer so as to form the first preform (PI) and obtaining several hollow impressions (EM) on one face of said first preform (PI), each of said hollow impressions (EM) being configured to delimit one of said volumes three-dimensional (112);

- placement in said mold (30) of a second layer of the second material (B) so as to cover said face of the first preform (PI) and fill said hollow indentations (EM) to form the three-dimensional volumes (112) ;

- compacting of the second layer so as to form the second preforms (P2);

- Assembly of the first preform (PI) and of the second preforms (P2) by sintering so as to obtain the composite part (11);

- Extraction from the mold (30) of said composite part (11);

- rectification of the second preforms (P2) consisting in removing the part of the second preforms (P2) located above said face of the first preform (PI) without removing the other part of the second preforms (P2) each forming a three-dimensional volume disposed in the first preform (PI).

6. Manufacturing process according to one of claims 1 to 5, characterized in that the sintering of the first preform (PI) and of the second preforms (P2) consists of conventional sintering, such as natural sintering, sintering under load (HP) or isostatic sintering, or in unconventional sintering, such as flash sintering (SPS), laser sintering or microwave sintering.

7. Composite part (11) obtained by the manufacturing process according to one of claims 1 to 6 comprising:

a substrate (111) made of a first material (A), said first material (A) being a ceramic or a ceramic matrix composite, and a plurality of three-dimensional volumes (112) disposed in said substrate (111), each of said three-dimensional volumes (112) being made of a second material (B) which is a ceramic or a ceramic matrix composite having electrical conductivity, permittivity and predefined thermal conductivity,

said three-dimensional volumes (112) being made up of the same ceramic or of the same ceramic matrix composite or else of different ceramics or ceramic matrix composites.

8. Composite part (11) according to claim 7, characterized in that said three-dimensional volumes (112) are arranged on one face of said substrate (111).

9. Composite part (11) according to claim 7, characterized in that said three-dimensional volumes (112) are arranged within said substrate (111).

10. Composite part (11) according to one of claims 7 to 9, characterized in that the first material (A) is composed of a mixture of AIN with Xi% m Y ₂ 0 ₃ and x ₂ % m CaF ₂ , the second material (B) being composed of a mixture of AIN with Xi% m Y ₂ 0 ₃ and x ₂ % m CaF ₂ and y vol.% MCG (multilayer of graphene) or of another carbonaceous material, or of a metal disulfide such as MoS ₂ , MXS ₂ WS ₂ with 0.5 <Xi <6 m%, 1 <x ₂ <7 m% and 0.01 <y <10 vol.%.

11. Composite part (11) according to one of claims 7 to 9, characterized in that the first material (A) is composed of a mixture of AlN and sintering additions and the second material (B) is composed:

of a mixture of AlN and y vol. % of graphene multilayers, or

of a mixture of AlN and y vol. % of carbon nanotubes, or

of a mixture of AlN and y vol. % of carbon,

with 0.01 <y <10 vol.%.

12. Composite part (11) according to one of claims 7 to 9, characterized in that the first material (A) is composed of Al ₂ 0 ₃ and the second material (B) is composed:

- a mixture of AI ₂ 0 ₃ and y vol. % of a metal disulfide and y vol. % of graphene multilayers, or

- a mixture of AI ₂ 0 ₃ and y vol. % of a metal disulfide and y vol. % of carbon nanotubes,

with 0.01 <y <10 vol.%.

13. Composite part (11) according to one of claims 7 to 9, characterized in that the first material (A) is composed of Si ₃ N ₄ or of SiAION and the second material (B) is composed:

of a mixture of Si ₃ N ₄ or SiAION respectively, and y vol. % of a metal disulfide and y vol. % of graphene multilayers, or

of a mixture of Si ₃ N ₄ or SiAION respectively, and y vol. % of a metal disulfide and y vol. % of carbon nanotubes,

with £ 0.01 y £ 10 vol.%.

14. An electrical component comprising a composite part (11) according to one of claims 7 to 13.

15. Electrical component according to claim 14 taking the form of an electronic power module, said composite part (11) constituting the insulating substrate of said electronic power module, said three-dimensional volumes (112) arranged in said substrate (111) of said. composite part (11) having an electrical conductivity and a permittivity different from those of said substrate (111) so as to reduce the electrical stress in the insulating substrate.

16. Electrical component according to claim 15, characterized in that said three-dimensional volumes (112) have an electrical conductivity in the plane perpendicular to the pressing rate greater than the electrical conductivity in the plane parallel to the pressing rate.

17. Electrical component according to claim 16, characterized in that the ratio of electrical conductivities, called anisotropy factor, in the planes parallel and perpendicular to the pressing axis is between 25 and 10 ¹² .