Method for controlling the quality of a composite structure comprising a thin c-SiC layer and composite structure

Abstract

This invention relates to a method for controlling the quality of a composite structure comprising a thin film made of monocrystalline silicon carbide disposed on a carrier substrate made of polycrystalline silicon carbide, the method comprising: a) examining the free surface (10a) of the thin film (10) using a combination of confocal visible light microscopy and photoluminescence imaging, thereby enabling the detection of defects referred to as secondary defects; b) performing preliminary identification of secondary defects based on visible light images by means of similarity using an image recognition algorithm, the image recognition algorithm having been trained for various defect types such as holes, bubbles, scratches, and crystal source defects; after completing step b), each Secondary defects are all associated with an identified defect type and have a certain level of similarity; c) At least some secondary defects are finally classified by applying the following first set of conditions: - If the similarity level associated with a secondary defect is greater than a higher level, the secondary defect is finally classified as an identified defect type; - If the similarity level associated with a secondary defect is between a low level and a high level, the photoluminescence image of the defect is analyzed; if a secondary defect is associated with a labeled PL defect type, the secondary defect is finally classified as an identified defect type; - In other cases, the secondary defect is finally classified as a non-defect.

Description

Technical Field

[0001] This invention relates to the field of semiconductor materials. Specifically, this invention relates to a method for inspecting the quality of a composite structure comprising a thin film made of monocrystalline silicon carbide disposed on a carrier substrate made of polycrystalline silicon carbide. Background Technology

[0002] Silicon carbide is a material particularly suitable for manufacturing power and radio frequency devices, even those that operate at very high temperatures.

[0003] Over the past decade, the quality of single-crystal silicon carbide (c-SiC) substrates has significantly improved, and simultaneously, the recognition and detection of various types of crystal defects that may exist in this material have become increasingly precise. JEITA (Japan Electronics and Information Technology Industries Association) standards comprise four documents covering the taxonomy of defects on / within thin films prepared by homoepitaxial growth on 4H-SiC substrates (EDR 4712 / 100) and non-destructive procedures for optical inspection of these defects (EDR 4712 / 200, / 300, / 400). In particular, a procedure for evaluating and referencing defects by combining optical inspection and photoluminescence imaging is provided.

[0004] Commercially available equipment, such as Lasertec’s SICA88, can be used in combination with visible light Nomarski prism confocal microscopy and photoluminescence technology (as described in T. Kimoto et al. [1], “Fundamentals of Silicon Carbide Technology Growth Characterization Devices and Applications”, p. 126, or D. Baierhofer et al. [2], “Materials Science in Semiconductor processing 140”, (2022) 106414), and is often used to inspect c-SiC substrates before or after epitaxy.

[0005] For example, Das et al. [3] (“Statistical analysis of killer and non-killer defects in SiC and the impact of device performance”, Material Science Forum, ISSN 1662-9752, Vol. 1004, pp. 458-463 (2020), Trans Tech Publications Ltd) proposed a method for statistical analysis of “killer” defects in devices fabricated on homoepitaxial films made of c-SiC and showed optical and photoluminescence images of typical c-SiC defects.

[0006] Despite the rapid development of high-quality c-SiC substrates, their price remains high and they are difficult to supply in large quantities. Therefore, it is advantageous to use thin-film transfer methods to fabricate composite structures comprising a thin film of single-crystal SiC (derived from a high-quality c-SiC donor substrate) on a low-cost carrier substrate (e.g., made of polycrystalline SiC (p-SiC)), which also offers advantages in conductivity. A well-known thin-film transfer method is Smart Cut. TM The method is based on light ion implantation and direct bonding assembly between a donor substrate made of c-SiC and a carrier substrate at a bonding interface. Implantation creates a fragile buried surface, along which separation occurs, thereby transferring a thin film made of c-SiC onto the carrier substrate to form a composite structure. This allows for the recovery and recycling of any remaining donor substrate, potentially enabling one or more transfers of other films. Epitaxy can then be performed on the thin film of the composite structure, followed by the fabrication of electronic devices.

[0007] For film transfer methods to be economically feasible, it is crucial to inspect the quality of the donor substrate to avoid transferring substandard films or introducing fatal defects into the epitaxial film. Therefore, it is necessary to detect and accurately classify crystal defects (referred to as primary defects) present on the donor substrate, allowing for the removal of substrates from the list that cannot produce films of the desired quality. Defect taxonomy in c-SiC is relatively well-known, and much research has focused on determining the types and sizes of defects in c-SiC epitaxial films that lead to component failure; thus, standards for classifying primary defects can be established for the transfer of c-SiC films.

[0008] Furthermore, the detection and identification of defects (referred to as secondary defects) on and / or within composite films (derived from donor substrates) made of polycrystalline SiC (p-SiC) substrates are highly complex because the p-SiC grains are clearly visible beneath the film. Therefore, on the one hand, an effective method is needed to inspect film quality, preventing epitaxial steps from continuing even when secondary defects in the film could result in substandard, fatal defect densities in the homogeneous epitaxial film; on the other hand, it is crucial that this inspection method can distinguish between defects in the film and potential “false defects” (or measurement noise) associated with the underlying grains in the p-SiC substrate.

[0009] Given the increasing application of film transfer technology in the electronic device manufacturing chain on c-SiC, the main need is to determine inspection methods that can reliably test composite structures.

[0010] The subject of this invention This invention solves the aforementioned problems. The present invention relates to a method for inspecting the quality of a composite structure comprising a thin film (derived from a donor substrate made of single-crystal silicon carbide) disposed on a carrier substrate made of polycrystalline silicon carbide, the inspection method being capable of reliably classifying any secondary defects present on and / or within the thin film. Summary of the Invention

[0011] This invention relates to a method for inspecting the quality of a composite structure comprising a thin film made of monocrystalline silicon carbide disposed on a carrier substrate made of polycrystalline silicon carbide, the method comprising the following steps: a) Using a combination of confocal visible light microscopy and photoluminescence imaging, the free surface of the thin film is examined to detect defects (called secondary defects) present on and / or within the film, and visible light and photoluminescence images are generated for each defect; b) Based on visible light images, secondary defects are initially identified by an image recognition algorithm through similarity. The image recognition algorithm has been trained for various defect types that may exist on and / or within the thin film, such as holes, bubbles, scratches, and crystal source defects. After completing step b), each secondary defect is associated with the identified defect type and has a certain level of similarity. c) Final classification of at least some secondary defects by applying the following first set of conditions: - If the similarity level associated with a secondary defect is greater than the higher level, the secondary defect is ultimately classified as an identified defect type; - If the similarity level associated with a secondary defect is between low and high, the photoluminescence image of the defect is analyzed using an image recognition algorithm trained for various labeled defect types; if the secondary defect is associated with a labeled defect type, the secondary defect is ultimately classified as the identified defect type. - In other cases, secondary defects are ultimately classified as non-defects.

[0012] Other advantageous and non-limiting features according to the invention, individually or in any technically feasible combination: The first set of conditions applies to secondary defects that are associated with holes or crystal source defects after step b) is completed; The high-level (referred to as the first high-level) and / or low-level (referred to as the first low-level) secondary defects used in step c) to associate with the hole in step b) are different from the high-level (referred to as the second high-level) and / or low-level (referred to as the second low-level) secondary defects used in step c) to associate with the crystal source defect in step b). The final classification step c) is also based on applying the following second set of conditions: - If a secondary defect is identified as a bubble-type defect in step b), and its similarity level is greater than that of the third level, then the secondary defect is ultimately classified as a bubble; - If a secondary defect is identified as a bubble-type defect in step b), and its similarity level is less than or equal to the third level, then the secondary defect is ultimately classified as a non-defect; The final classification step c) is also based on applying the following third set of conditions: - If a secondary defect is identified as a scratch-type defect in step b), and its similarity level is greater than that of the fourth level, then the secondary defect is ultimately classified as a scratch; - If a secondary defect is identified as a scratch-type defect in step b), and its similarity level is less than or equal to level four, then the secondary defect is ultimately classified as a non-defect; The size of each secondary defect detected in step a) is greater than 1 micrometer, 5 micrometers, or even 10 micrometers; Step c) results in the grading of composite structures and the separation of de-categorized composite structures from compliant composite structures through physical sorting. When the density of crystal source defects in a composite structure is greater than 0.25 defects / cm² and / or the total density of bubbles and pores is greater than 0.2 defects / cm², the composite structure is declassified.

[0013] The present invention also relates to a composite structure formed by a thin film made of monocrystalline silicon carbide disposed on a carrier substrate made of polycrystalline silicon carbide, wherein the density of crystal source defects in the composite structure is less than or equal to 0.25 defects / cm², and the total density of bubbles and pores is less than or equal to 0.2 defects / cm². Attached Figure Description

[0014] Further features and advantages of the present invention will become apparent from the following detailed description of the invention, taken in conjunction with the accompanying drawings, wherein: [ Figure 1 ] Figure 1 The composite structures with donor substrate, no epitaxial film, and epitaxial film are shown; [ Figure 2a ] [ Figure 2b ] Figure 2a and Figure 2b Various defect types and their final classifications are shown, obtained under specific conditions by applying the inspection method according to the present invention. Detailed Implementation

[0015] In the following description, "first-order defect" generally refers to a defect present on and / or within the donor substrate 1; "second-order defect" generally refers to a defect present on and / or within the thin film 10 of the composite structure 100; and "third-order defect" generally refers to a defect present on and / or within the epitaxial film 150 grown on the thin film of the composite structure 100.

[0016] Within the principal plane (x, y), the donor substrate 1 and the composite structure 100 (with or without the epitaxial film 150) are preferably in the form of circular wafers with diameters of 100 mm, 150 mm, 200 mm, or even larger. However, they can also take any other form for subsequent processing to manufacture components. The thickness of the substrate and structure is along... Figure 1 The z-axis extension is shown in the figure. The front surfaces 1a, 10a, and 150a of the substrate 1 and structure 100 are inspected, and the aforementioned defects may exist on these surfaces.

[0017] Before beginning the fabrication of the composite structure 100 (which includes transferring a thin film made of monocrystalline silicon carbide (c-SiC) onto a carrier substrate made of polycrystalline silicon carbide (p-SiC), it is important to check the quality of the donor substrate 1 (made of, for example, 4H-SiC type c-SiC), from which the thin film will be obtained.

[0018] Preferably, the front side 1a of the donor substrate 1 is inspected using photoluminescence imaging technology (as described in references [1] and [2] cited in the introduction) to detect and classify any defects (hereinafter referred to as first-order defects).

[0019] First-order defects were revealed using photoluminescence imaging because the intensity of the photoluminescence (PL) signal emitted from the front side 1a changed locally after excitation with the incident beam. This localized change in PL signal intensity reflects alterations in material properties (stress, roughness, smoothness, etc.), corresponding to first-order defects. Depending on the characteristics of the defect, it may appear as a white or black spot in the generated PL image.

[0020] This inspection step can be performed on known equipment such as the SICA88 (from Lasertec), a photoluminescence scanner (from Intego), or a MiPlatoSiC (from EtaMax). This equipment typically combines optical microscopy and photoluminescence images of the same defect, thereby establishing size (size, surface area) and intensity (contrast) criteria based on the measured optical microscopy and photoluminescence signals; it also establishes classification criteria (predefined defect categories) for each primary defect based on a learning algorithm. These predefined categories typically correspond to defects such as microtubules, stacking faults (SF), pits, bumps, scratches, and particles.

[0021] After inspecting and potentially declassifying donor substrates 1 that have defect densities and / or types incompatible with the available films for transfer, the composite structure 100 can be fabricated. Any thin-film transfer technique can be used, particularly the Smart Cut method. This method, based on light ion implantation and direct assembly via molecular adhesion, is well-known and will not be elaborated upon herein. To obtain the composite structure 100, a donor substrate 1 made of c-SiC and a carrier substrate 20 made of p-SiC are used.

[0022] Typically, the thickness of the thin film 10 ranges from 20 nanometers to 1,000 nanometers, and the roughness of its free surface is less than or equal to 0.5 nanometer RMS, or even less than or equal to 0.1 nanometer RMS (e.g., measured by atomic force microscopy (AFM) in a scanning range of 10 x 10 square micrometers to 30 x 30 square micrometers). The thickness of the carrier substrate 20 can range from about 50 micrometers to several hundred micrometers, for example, between 200 micrometers and 700 micrometers. Optionally, the composite structure 100 may include an intermediate film between the thin film 10 and the carrier substrate 20. This intermediate film may be made of a dielectric material, a semiconductor material, or a metallic material (e.g., silicon oxide, silicon, silicon carbide, tungsten, titanium, etc.). A bonding interface exists between the thin film 10 and the carrier substrate 20, or near the intermediate film.

[0023] This invention relates to a method for inspecting the quality of a composite structure 100, the composite structure 100 comprising a thin film 10 made of c-SiC disposed on a carrier substrate 20 made of p-SiC. Figure 1 ).

[0024] The inspection method includes the following step a): using a combination of visible light confocal microscopy (differential interference contrast-DIC) and photoluminescence imaging to inspect the free surface 10a of the thin film 10, thereby detecting defects (referred to as secondary defects) present on and / or within the thin film 10. The size (or equivalent diameter) of the detected defects is typically greater than or equal to 1 micrometer, greater than or equal to 5 micrometers, or even greater than or equal to 10 micrometers.

[0025] Differential interference contrast (also known as Nomarski) is an illumination technique that uses the interference of light waves to identify minute changes in surface morphology.

[0026] Photoluminescence imaging is based on the localized changes in the intensity of the photoluminescence (PL) signal emitted from a free surface 10a after excitation by an incident electromagnetic beam. The remainder of this specification will preferentially use the SICA88 device, where the incident excitation beam has a wavelength of 313 nm, and the emitted photoluminescence signal is collected within a wavelength range of 700 nm to 1,000 nm. Localized changes in the PL signal intensity reflect alterations in material properties (stress, roughness, smoothness, etc.), corresponding to second-order defects. Depending on the characteristics of the defect, it may appear as a white or black spot on the generated PL image.

[0027] In step a), for each detected secondary defect, a visible light image (hereinafter referred to as the DIC image) and a photoluminescence image (hereinafter referred to as the PL image) are generated.

[0028] Secondary defects may be caused by primary defects or by particles or other surface contaminants that were not completely removed before assembling the donor substrate 1 and the carrier substrate 20. Therefore, these defects may be crystal source defects, bonding defects (e.g., pores, i.e., local missing parts of the film 10) or bubbles (defects at the bonding interface 40 where the film 10 is not bonded and forms bubbles), or even surface defects (e.g., scratches or peeling fragments) present on the front side 10a of the film.

[0029] The inspection method then includes step b), which is to perform preliminary identification of secondary defects based on similarity using DIC images.

[0030] The aforementioned devices (especially the SICA88) are capable of identifying secondary defects based on images captured by a visible light microscope using image recognition algorithms. The algorithm is trained by providing validated images of potential defects (e.g., pores, bubbles, scratches, fragments, crystal source defects) that may exist on and / or within the thin film 10; thus, the algorithm learns to identify each of these defects and calculates the similarity level between the DIC image of the detected defect and its database, taking into account factors such as size, shape, color, and contrast criteria.

[0031] After completing step b), each secondary defect is associated with an identified defect type (e.g., hole, bubble, scratch, fragment, or crystal source defect) and has a certain similarity level (ranging from 0 to 1).

[0032] Because a polycrystalline carrier substrate 20 exists beneath the thin film 10, the initial identification automatically performed by the device is insufficient to reliably determine the quality of the composite structure 100. In fact, such identification is affected by the underlying grains, and the similarity level assigned to secondary defects does not itself determine whether the defect actually belongs to the identified type, or whether it is a "false defect," such as a defect associated with p-SiC grains.

[0033] Therefore, the inspection method includes step c), which is to finally classify any secondary defects by applying the following specific conditions.

[0034] The first set of conditions applies specifically to secondary defects identified as pores or crystal source defects in step b): (i) If the similarity level associated with a secondary defect is greater than that of a higher level, the secondary defect is ultimately classified as an identified defect type; (ii) If the similarity level associated with a secondary defect is between low and high, the PL image of the defect is analyzed using an image recognition algorithm trained for various defect types. Two scenarios may then occur: - If a secondary defect is associated with an already labeled defect type, the secondary defect is ultimately classified as an identified defect type; - If a secondary defect is not identified as a labeled defect type, the secondary defect will ultimately be classified as a non-defect (i.e., considered a "false defect"). (iii) Finally, if the similarity level associated with a secondary defect is lower than that of a lower-level defect, the secondary defect is ultimately classified as a non-defect.

[0035] A “false defect” refers to a defect detected in step b) that is associated with a given type (with a certain level of similarity) but does not actually correspond to the type: it may correspond to another type of defect in the film, a measurement artifact, or a grain image in the underlying carrier substrate 20.

[0036] The PL image recognition algorithm is trained by providing verified PL images of defects that may exist on and / or within the thin film 10. The algorithm learns to identify each of these defects and categorizes them into predefined classes (i.e., associated with labeled defect types) taking into account size, shape, contrast criteria, etc. Thus, the labeled defect types correspond to real defects whose features on the PL image are reflected by specific criteria (such as those mentioned above).

[0037] In condition (ii), when the PL image associated with the analyzed secondary defect corresponds to one of the labeled defect types, this confirms that it is a “real defect” and the secondary defect is therefore classified into the type identified in step b); however, when the PL image does not correspond to any labeled defect type and / or the PL image does not exhibit any specific contrast (no PL signal), the secondary defect is considered a “false defect” even if it is associated with an identified defect type in step b).

[0038] It should be noted that the high level (referred to as the first high level) and / or low level (referred to as the first low level) of the similarity level of the DIC image used in step c) to determine the secondary defects associated with the hole are different from the high level (referred to as the second high level) and / or low level (referred to as the second low level) used for the secondary defects associated with the crystal source defects.

[0039] The second set of conditions can be applied to secondary defects identified as bubbles in step b): (iv) If the similarity level associated with a secondary defect identified as a bubble is greater than the third level, then the secondary defect is ultimately classified as a bubble; (v) If the similarity level is less than or equal to the third level, the secondary defect is ultimately classified as a non-defect.

[0040] The third set of conditions can be applied to secondary defects identified as scratches in step b): (vi) If the similarity level associated with a secondary defect identified as a scratch is greater than the fourth level, the secondary defect is ultimately classified as a scratch; (vii) If the similarity level is less than or equal to the fourth level, the secondary defect is ultimately classified as a non-defect.

[0041] This final classification step c) can reliably classify serious defects present on and / or within the film 10 and avoid considering “false defects,” particularly those related to the underlying grains of the test carrier substrate 20.

[0042] For example, this inspection method is applied to a composite structure 100 comprising a 600 nm thick film made of c-SiC disposed on a p-SiC carrier substrate with a diameter of 150 mm. The equipment used is an SICA88. Figure 2a and Figure 2b In the illustrated embodiment, the first low level and the first high level of similarity used in step c) to determine the association of secondary defects identified as holes are set to 0.3 and 0.85, respectively. The second low level and the second high level of similarity used in step c) to determine the association of secondary defects identified as crystal source defects are set to 0.3 and 0.85, respectively. The third level of similarity used in step c) to determine the association of secondary defects identified as bubbles is set to 0.85. Finally, the fourth level of similarity used in step c) to determine the association of secondary defects identified as scratches is set to 0.97.

[0043] Figure 2a and Figure 2b Defects of the types of holes, crystal sources, bubbles, and scratches detected on and / or within the thin film 10 are shown.

[0044] After completing step c) of the inspection method, a grade can be assigned to the inspected composite structure 100. Specifically, a higher quality grade can be assigned to structures 100 with a total density of bubbles and pores less than or equal to 0.2 defects / cm². Preferably, this grade is assigned to composite structures 100 with a total density of bubbles, pores, and scratches less than or equal to 0.2 defects / cm². Even more preferably, an excellent quality grade can be assigned to structures 100 with a total defect density (excluding crystal source defects) less than or equal to 0.2 defects / cm².

[0045] The acceptable density of crystal source defects is defined based on the quality level of the donor substrate 1 from which the thin film 10 originates. Advantageously, a higher quality level can be assigned to a structure 100 where the density of crystal source defects is less than or equal to 0.25 defects / square centimeter.

[0046] Conversely, if the total density of defects is greater than the aforementioned threshold, it can be defined as a lower grade, causing the composite structure 100 to be declassified after completing step c).

[0047] Therefore, the present invention also relates to a composite structure 100 formed by a thin film 10 made of monocrystalline silicon carbide disposed on a carrier substrate 20 made of polycrystalline silicon carbide. The composite structure is ultimately inspected using the inspection method according to the present invention, and the composite structure has: - The density of crystal source defects is less than or equal to 0.25 defects / cm², less than or equal to 0.2 defects / cm², even less than or equal to 0.18 defects / cm², even less than or equal to 0.15 defects / cm²; and - The total density of bubbles and holes (and optional scratches) is less than or equal to 0.2 defects / cm², less than or equal to 0.18 defects / cm², less than or equal to 0.15 defects / cm², or even less than or equal to 0.12 defects / cm², or even less than or equal to 0.1 defects / cm².

[0048] After grading, the composite structures 100 are physically sorted to separate the unclassified structures 100 from the compliant structures 100.

[0049] In subsequent manufacturing steps, SiC can be epitaxially grown on the thin film 10 of the conforming composite structure 100 to form an epitaxial film 150, on which the device will be manufactured (and within the epitaxial film). Figure 1 The epitaxial growth step can be performed using techniques known in the prior art.

[0050] Prior to the epitaxial step, the method for inspecting the quality of the composite structure 100 ensures that the density of level 3 fatal defects related to the quality of the thin film 10 is controlled, even below the given specifications. This is because the method is able to reliably classify any defects present on and / or within the thin film 10 without declassifying the expensive composite structure 100 based on “false defects”, especially those caused by the underlying polycrystalline substrate, thereby avoiding a decline in quality.

[0051] Of course, the present invention is not limited to the described embodiments and examples, and alternative embodiments may be applied without departing from the scope of the invention as defined by the claims.

Claims

1. A method for inspecting the quality of a composite structure (100), the composite structure comprising a thin film (10) made of monocrystalline silicon carbide disposed on a carrier substrate (20) made of polycrystalline silicon carbide, the method comprising the steps of: a) Using a combination of confocal visible light microscopy and photoluminescence imaging, the free surface (10a) of the film (10) is examined to detect defects known as secondary defects that exist on and / or within the film (10), and visible light and photoluminescence images are formed for each defect; b) Based on visible light images, secondary defects are initially identified by image recognition algorithms through similarity. The image recognition algorithms have been trained for various defect types that may exist on and / or within the thin film (10), such as holes, bubbles, scratches, and crystal source defects. After completing step b), each secondary defect is associated with the identified defect type and has a certain similarity level. c) Final classification of at least some secondary defects by applying the following first set of conditions: - If the similarity level associated with a secondary defect is greater than the higher level, the secondary defect is ultimately classified as an identified defect type; - If the similarity level associated with a secondary defect is between low and high, the photoluminescence image of the defect is analyzed using an image recognition algorithm trained for various defect types. If a secondary defect is associated with a labeled defect type, the secondary defect is ultimately classified as an identified defect type. - In other cases, secondary defects are ultimately classified as non-defects.

2. The method for inspecting the quality of a composite structure (100) according to claim 1, wherein, The first set of conditions applies to secondary defects that are associated with holes or crystal source defects after step b).

3. The method for inspecting the quality of a composite structure (100) according to claim 2, wherein, The high-level term, referred to as the first high-level and / or the low-level term, used in step c) for secondary defects associated with the hole in step b), is different from the high-level term, referred to as the second high-level and / or the low-level term, referred to as the second low-level term, used in step c) for secondary defects associated with the crystal source defect in step b).

4. The method for inspecting the quality of a composite structure (100) according to any one of claims 1 to 3, wherein, The final classification step c) is also based on applying the following second set of conditions: - If a secondary defect is identified as a bubble-type defect in step b), and its similarity level is greater than that of the third level, then the secondary defect is ultimately classified as a bubble; - If a secondary defect is identified as a bubble-type defect in step b), and its similarity level is less than or equal to the third level, then the secondary defect is ultimately classified as a non-defect.

5. The method for inspecting the quality of a composite structure (100) according to any one of claims 1 to 4, wherein, The final classification step c) is also based on applying the following third set of conditions: - If a secondary defect is identified as a scratch-type defect in step b), and its similarity level is greater than that of the fourth level, then the secondary defect is ultimately classified as a scratch; - If a secondary defect is identified as a scratch-type defect in step b) and its similarity level is less than or equal to the fourth level, then the secondary defect is ultimately classified as a non-defect.

6. The method for inspecting the quality of a composite structure (100) according to any one of claims 1 to 5, wherein, The size of each secondary defect detected in step a) is greater than 1 micrometer, 5 micrometers or even 10 micrometers.

7. The method for inspecting the quality of a composite structure (100) according to any one of claims 1 to 6, wherein, Step c) results in the grading of composite structures (100) and the separation of de-categorized composite structures from compliant composite structures by physical sorting.

8. The method for inspecting the quality of a composite structure (100) according to claim 7, wherein, The composite structure (100) is declassified when the density of crystal source defects in the composite structure (100) is greater than 0.25 defects / cm² and / or the total density of bubbles and pores is greater than 0.2 defects / cm².

9. A composite structure (100) formed from a thin film (10) made of monocrystalline silicon carbide disposed on a carrier substrate (20) made of polycrystalline silicon carbide, wherein the density of crystal source defects of the composite structure (100) is less than or equal to 0.25 defects / cm² and the total density of bubbles and pores is less than or equal to 0.2 defects / cm².

Images