High-quality silicon carbide substrate

Abstract

This silicon carbide substrate has high crystal quality and belongs to the field of silicon carbide production and processing technology. The silicon carbide substrate includes a first main surface and a second main surface. The first main surface has a central region and an annular region surrounding the central region. The annular region extends inward from the edge of the substrate and has a width of 5 to 30 mm. The central region is divided into square regions, each with a side length of 5 mm. The internal stress in each square region is smaller than the internal stress in the annular region. The internal stress is the stress value detected at a location extending at least 30 μm perpendicularly into the silicon carbide substrate from the first or second main surface. The internal stress in the central region of the silicon carbide substrate is relatively low, and the stress on the substrate is uniformly distributed. This demonstrates that the quality of the silicon carbide substrate is relatively high, which can improve the quality of subsequent crystals and thereby expand the range of applications for the silicon carbide substrate.

Description

[Technical Field]

[0001] This application relates to a silicon carbide substrate with high crystal quality and belongs to the field of silicon carbide production and processing technology. [Background technology]

[0002] Silicon carbide single crystal substrates possess excellent physical properties, including a large bandgap, high resistivity and thermal conductivity, and high dielectric breakdown field strength, making them suitable semiconductor materials for manufacturing GaN-based high-frequency microwave devices. With the continued development of 5G technology, the demand for commercially available silicon carbide single crystal substrates continues to expand, and more importantly, large-scale commercial applications demand higher quality from silicon carbide single crystal substrates.

[0003] During the growth process, silicon carbide crystals are prone to introducing various defects such as polycrystalline, polymorphic, microtubule, and dislocation defects due to Si / C ratio imbalances, impurity introduction, and changes in the potential temperature gradient. These defects cause distortion of the crystal lattice, and in more severe cases, alter the crystal type, resulting in polymorphic coexistence such as 4H, 6H, 3C, and 15R. This distortion of the crystal lattice releases stress due to the occurrence of defects, thus introducing internal stress into the silicon carbide. This degrades the quality of the silicon carbide crystal, limits the range of use, and reduces product yield. While the quality of internal stress in silicon carbide substrates is better depending on the temperature field or the presence of defects, a phenomenon called stress concentration exists in certain regions. This results in a non-uniform stress distribution in the resulting silicon carbide substrate, which affects the production and processing of downstream products. When used for growing crystals or epitaxial layers, this is inherited by the crystals or epitaxial layers, causing stress concentration in those layers. [Overview of the project] [Problems that the invention aims to solve]

[0004] To solve the above problem, a high-quality silicon carbide substrate is provided, which is divided into a central region and an annular region. The central region can be further divided into multiple square regions with side lengths of 5 mm, which not only satisfies detection accuracy but also allows for region division of the substrate, enabling quantitative analysis of whether or not stress concentration exists on the substrate. The internal stress in each square region is smaller than that in the annular region, demonstrating that the internal stress in the central region of the silicon carbide substrate is relatively low and that the stress on the substrate is uniformly distributed, thus proving the relatively high quality of the silicon carbide substrate and expanding the range of applications. The silicon carbide substrate can be used to obtain epitaxial wafers through processing or as a seed crystal for crystal growth, improving the quality of subsequent crystals, thereby reducing stress in the epitaxial wafer or crystal, and thereby obtaining a high-quality epitaxial wafer or crystal. [Means for solving the problem]

[0005] According to one aspect of the present application, a silicon carbide substrate of high crystal quality is provided, wherein the silicon carbide substrate has a diameter of 150 mm or more, and the silicon carbide substrate includes a first main surface and a second main surface. The first main surface has a central region and an annular region surrounding the central region, and the annular region extends inward from the edge of the substrate with a width of 5 to 30 mm. The central region is divided into square regions, each having a side length of 5 mm. The internal stress in each of the square regions is smaller than the internal stress in the annular region, and the internal stress is the stress value detected at a location extending at least 30 μm perpendicularly into the silicon carbide substrate from the first or second main surface.

[0006] The internal stress in the annular region is greater than the internal stress in each square region in the central region. This proves that the smaller the width of the annular region extending inward from the edge of the substrate, the larger the area of ​​the central region, and therefore the lower the internal stress of the silicon carbide substrate. If the internal stress in the annular region is too high, it is possible to remove some of the crystals in the annular region during actual use, retaining only the silicon carbide crystals in the central region, thereby obtaining a higher quality silicon carbide substrate.

[0007] Optionally, in either the plane parallel to the first principal surface and / or the second principal surface, the radial internal stress in the square region is -10 to 10 MPa, and the radial internal stress in the annular region is 5 to 15 MPa.

[0008] Preferably, the radial internal stress in the square region is -9.4 to 9.2 MPa, and the radial internal stress in the annular region is 5 to 14.7 MPa.

[0009] When a silicon carbide substrate is subjected to tensile and compressive stress, the pitch d of the crystal planes elongates and contracts accordingly. When testing the stress using Raman arithmetic, the Raman peak intensity shifts to either a low or high frequency. When the substrate is subjected to tensile stress, the Raman peak shifts to a low frequency, resulting in a positive stress value. When the substrate is subjected to compressive stress, the Raman peak shifts to a high frequency, resulting in a negative stress value. Therefore, in this application, the sign before the stress value indicates the direction of force within the substrate, and the absolute value of the value indicates the magnitude of the stress. For example, if the radial stress in a square region of a silicon carbide substrate is -5 MPa and the radial internal stress in an annular region is 6 MPa, this indicates that the radial internal stress in the square region is compressive stress, the radial internal stress in the annular region is tensile stress, and the radial internal pressure in the square region is smaller than the radial internal pressure in the first annular region.

[0010] As can be seen from the radial in-body stress values in the above-mentioned square region and annular region, compressive stress and tensile stress exist in the square region, and the annular region is a tensile stress region. It is proved that the closer the numerical value of the square region is to the annular region, the more uniform the distribution of the radial in-body stress of the silicon carbide substrate is.

[0011] Optionally, in any plane parallel to the first major surface and / or the second major surface, S max1 represents the maximum value of the absolute value of the radial in-body stress in all the above-mentioned square regions, and S max2 represents the maximum value of the absolute value of the radial in-body stress in the annular region, and 1.8 MPa ≤ S max2 - S max1 ≤ 5.2 MPa.

[0012] Optionally, in any plane parallel to the first major surface and / or the second major surface, S1 represents the average value of the absolute value of the radial in-body stress in all the above-mentioned square regions, S2 represents the average value of the absolute value of the radial in-body stress in any of the above-mentioned square regions, and -5 MPa ≤ S1 - S2 ≤ 5 MPa.

[0013] Perform stress multi-point detection on the silicon carbide substrate. The above-mentioned average value of the absolute value is a numerical value obtained by dividing the sum of the absolute values of the stresses at all detection points in the region by the number of detection points. The detection of the multiple points may be performed by selecting different points at multiple positions in the region, or a Mapping test may be performed on the substrate.

[0014] Preferably, -4.8 MPa ≤ S1 - S2 ≤ 4.9 MPa, and more preferably, -2.4 MPa ≤ S1 - S2 ≤ 3.3 MPa. The difference between S1 - S2 represents the difference between the stress of the entire central region and the stress of each square region. The smaller the difference, the more uniform the stress distribution of the central region is proved.

[0015] Optionally, in any plane parallel to the first major surface and / or the second major surface, S1 represents the average value of the absolute value of the radial internal stress in all of the square regions, S3 represents the average value of the absolute value of the radial internal stress in the annular region, and 0.3 ≤ S1 / S3 ≤ 0.94.

[0016] Preferably, 0.31 ≤ S1 / S3 ≤ 0.94. The ratio of S1 / S3 represents the difference in the radial internal stress between the central region and the annular region. The closer the ratio is to 1, the smaller the difference between the annular region and the central region, indicating that the distribution of the radial internal stress in the silicon carbide substrate is more uniform.

[0017] Optionally, in any plane parallel to the first major surface and / or the second major surface, S max3 represents the maximum value of the radial internal stress in any of the square regions, S min3 represents the minimum value of the radial stress in any of the square regions, ΔS1 = S max3 - S min3 and 0 MPa ≤ ΔS1 ≤ 10 MPa. In any plane parallel to the first major surface and / or the second major surface, S max4 represents the maximum value of the radial internal stress in the annular region, S min4 represents the minimum value of the radial stress in the annular region, ΔS2 = S max4 - S min4 and 0 MPa ≤ ΔS2 ≤ 5 MPa.

[0018] Preferably, 0 MPa ≤ ΔS1 ≤ 9.5 MPa and 0.1 MPa ≤ ΔS2 ≤ 4.9 MPa.

[0019] The maximum and minimum values ​​of the radial internal stress described above are the true values ​​of the radial internal stress, distinguishing between compressive and tensile stress. For example, in either the horizontal plane parallel to the first principal surface and / or the second principal surface, the radial internal stress in the square region is -5 to 2 MPa. Thus, the maximum value of the total radial stress is 2 MPa, the minimum value of the total radial stress is -5 MPa, and ΔS1 is 7 MPa. ΔS1 and ΔS2 represent the degree of change between the pitches of the radial crystal planes of the substrate at a microscopic level. The smaller ΔS1 and ΔS2 are, the smaller the change in the pitch of the radial crystal planes.

[0020] An axial test is performed by extending perpendicularly from any point on the first main surface into the silicon carbide substrate, and the axial internal stress in the square region is -10 to 10 MPa, and the axial internal stress in the annular region is -15 to 15 MPa.

[0021] Preferably, the internal stress in the axial direction of the square region is -5 to 5 MPa, and the internal stress in the axial direction of the annular region is -10 to 10 MPa.

[0022] Optionally, an axial test is performed by extending perpendicularly into the silicon carbide substrate from any point on the first main surface, S max5 Let S be denoted as the maximum value of the axial stress along the axis of the square region, min5 Let ΔS3 = S be the minimum value of the axial stress along the axis of the square region, and ΔS3 = S max5 -S min5 Therefore, 0MPa ≤ ΔS3 ≤ 10MPa, An axial test is performed by extending perpendicularly from any point on the first main surface into the silicon carbide substrate, S max6 Let S be denoted as the maximum axial stress in the annular region. min6 Let ΔS4 = S be the minimum value of axial stress in the annular region. max6 -S min6 Therefore, 0MPa ≤ ΔS4 ≤ 15MPa.

[0023] Preferably, 0.1 MPa ≤ ΔS3 ≤ 9.7 MPa and 0.3 MPa ≤ ΔS4 ≤ 14.3 MPa.

[0024] The meaning of the maximum and minimum values ​​of axial internal stress described above is the same as the meaning of the maximum and minimum values ​​of radial internal stress. By separating compressive and tensile stress in the axial direction, for example, in an arbitrary axis, the axial internal stress in a square region is -5 to 1 MPa. In this case, the maximum value of axial internal stress is 1 MPa, the minimum value of axial internal stress is -5 MPa, and ΔS3 is 6 MPa. ΔS3 and ΔS4 represent the degree of change between the pitches of the crystal planes in the axial direction of the substrate at a microscopic level. The smaller ΔS3 and ΔS4 are, the smaller the change in the pitch of the crystal planes in the axial direction of the substrate.

[0025] Optionally, S4 represents the average value of the absolute values ​​of the axial internal stress in all the square regions, and S5 represents the average value of the absolute values ​​of the radial internal stress in any of the square regions, such that -5MPa ≤ S4 - S5 ≤ 5MPa, and preferably -4.2MPa ≤ S4 - S5 ≤ 4.8MPa.

[0026] Optionally, S4 represents the average value of the absolute values ​​of the axial internal stress in all the square regions, and S6 represents the average value of the absolute values ​​of the axial internal stress in the annular region, with 0.1 ≤ S4 / S6 ≤ 0.8, preferably 0.1 ≤ S4 / S6 ≤ 0.79.

[0027] The stress values ​​described in this application are both absolute stress and relative stress. Absolute stress can reflect the difference between the substrate and a perfect, defect-free SiC crystal and is used to determine the stress level of the SiC substrate. However, since it is difficult to realize a perfect substrate without SiC with current technology, relative stress can be used to determine the relative stress distribution within the substrate surface.

[0028] For example, this explains why a substrate may have high absolute stress, low relative stress, and many in-plane defects. However, if the stress distribution is uniform, the crystal quality is poor, and the uniformity quality of each region of the substrate is also poor, while the absolute stress is low and the relative stress is high, the overall quality of the substrate is good, but there are localized abnormal quality variations.

[0029] In this application, the reference value for absolute stress is a reference stress calculated based on standard Raman peak bits obtained from SiC perfect crystal lattice parameters, and is denoted as 0. For relative stress, one reference value is taken for the entire test plane, and the relative stress value for each test point is obtained by performing a correlation calculation between the numerical value of each test point and the reference value. The selection of the reference value includes, but is not limited to, the median, mean, mode, or result obtained by calculating the median, mean, mode, or other statistical function of all stress values ​​across the entire test plane.

[0030] Optionally, the diameter of the silicon carbide substrate is 200 mm or more, and the width extending inward from the edge of the substrate in the annular region is 5 to 15 mm.

[0031] Optionally, the thickness of the silicon carbide substrate is 100 μm or more.

[0032] Optionally, the angle of deviation of the first principal surface and / or the second principal surface with respect to the {0001} plane is 4° or less. Preferably, the curvature of the silicon carbide substrate is -50 to 50 μm, more preferably -20 μm to 20 μm, and more preferably -10 μm to 10 μm.

[0033] The crystal type of the silicon carbide substrate is optionally one of 2H-SiC, 4H-SiC, 6H-SiC, 3C-SiC, and 15R-SiC, with 4H-SiC being preferred.

[0034] Optionally, the silicon carbide substrate may be a semi-insulating silicon carbide substrate or a conductive silicon carbide substrate.

[0035] Selectively, the local thickness variation in the central region is 0.1 to 2 μm, and the local thickness variation in the annular region is 0.1 to 2 μm.

[0036] A high-quality silicon carbide substrate described in any one of the above is manufactured by processing a silicon carbide single crystal, and a method for manufacturing a silicon carbide single crystal provided according to another aspect of this application is: Step (1) involves adding silicon carbide powder to a graphite crucible, 10 inside the reactor -6 The vacuum is reduced to below mbar, followed by the flow of high-purity inert gas up to 300-500 mbar. This process is repeated 2-3 times, and finally the furnace body is filled to 10 -6 Step (2) involves vacuuming to below mbar, Step (3) involves flowing high-purity inert gas into the furnace body, raising the pressure to 10-100 mbar within 1-3 hours, and maintaining the pressure by continuously flowing high-purity inert gas. In the crystal growth stage, if the pressure is kept constant, step (4) involves raising the temperature inside the furnace to the single crystal growth temperature of 2200K to 2800K within 3 to 5 hours, and the growth time being 30 to 150 hours. The process includes step (5) obtaining a silicon carbide single crystal after the growth of the single crystal is complete, opening the furnace body, and removing the graphite crucible.

[0037] The graphite crucible is optionally coated with a TaC coating layer, the TaC coating layer is deposited on the surface of the graphite substrate, through-holes are uniformly distributed in the TaC coating layer, and the density of the through-holes is 160-310 holes / mm². 2 That is the case.

[0038] This TaC coating layer contains through-holes, which improves the transmission rate of the silicon carbide raw material atmosphere and simultaneously improves the uniformity of the temperature field during crystal growth. This increases the crystal growth rate and reduces internal stress in the crystal, resulting in a high-quality, low-stress silicon carbide crystal.

[0039] Optionally, the diameter of the through-hole is 2 to 7 μm, preferably 2.3 to 6.1 μm. This diameter not only improves the transmission speed of the silicon carbide raw material atmosphere, but also does not reduce the protective performance for the graphite substrate, avoids the generation of edge stress on the crystal structure due to thermal expansion and pressing of the graphite substrate in a hot electric field, and further reduces stress in the crystal structure.

[0040] The thickness of the TaC coating layer is optionally 30 to 400 μm, the TaC crystal grains in the TaC coating layer are arranged in a stacked dislocation row manner, and the size of the TaC crystal grains is 15 to 50 μm. The TaC crystal grains in the TaC coating layer grow oriented along the {200} and {220} directions. This growth direction of the TaC crystal grains can improve the temperature resistance and thermal chemical erosion resistance of the TaC coating layer, thereby reducing interference with the crystal growth environment of the graphite substrate and improving the stability of the crystal environment. [Effects of the Invention]

[0041] The beneficial effects of this invention include, but are not limited to, the following:

[0042] 1. Residual stress exists within the silicon carbide crystal due to thermal stress during crystal growth and defects within the crystal. This residual stress is the internal stress of the silicon carbide substrate. The internal stress in the central region of the silicon carbide substrate of this application is smaller than the internal stress in the annular region, demonstrating the good quality of the silicon carbide substrate and providing an advantage in the production and processing of downstream products.

[0043] 2. The silicon carbide substrate of this invention has low radial internal stress in the annular region, and as can be seen from the stress values ​​of each square region in the annular and central regions, tensile and compressive stress can coexist in the annular region, while only tensile stress exists in the annular region. Therefore, the stress distribution of the silicon carbide substrate can be controlled within a reasonable range based on the differences in downstream products used in the manufacture of the silicon carbide substrate, thereby expanding the range of use of the silicon carbide substrate.

[0044] 3. The silicon carbide substrate of this application demonstrates that the distribution of radial stress and axial stress in each square region within the central region is uniform, indicating high uniformity of the silicon carbide substrate, and that the difference in radial stress between the central region and the annular region is small.

[0045] 4. This invention utilizes a TaC coating layer having through-holes formed on the crucible surface to ensure the transmission of the raw material atmosphere during crystal growth while simultaneously improving the uniformity of the temperature field. This reduces internal stress in the silicon carbide substrate and minimizes the difference between the central region and the annular region of the silicon carbide substrate.

[0046] 5. Based on the stress distribution information in the central and annular regions of the silicon carbide substrate of this invention, and the stress distribution information in each square region within the central region, the microscopic unit cell parameters of the substrate can be analyzed from the stress values. This provides a certain feedback effect, allowing for adjustment of the process parameters for crystal growth, thereby enabling the production of high-quality silicon carbide substrates.

[0047] 6. The quantitative test method of this invention further includes stresses that are not released due to defects in the crystal lattice present in the substrate, thereby improving the accuracy of stress detection in the silicon carbide substrate and reflecting all stress information in the central and annular regions of the entire silicon carbide substrate. [Brief explanation of the drawing]

[0048] The drawings described herein are provided to provide a further understanding of the present application and constitute part of the present application; the exemplary embodiments and descriptions thereof are used for interpretation of the present application and do not constitute an unreasonable limitation thereto. [Figure 1] This is a schematic diagram of the stress within the substrate during a Raman pitch focusing test and an isoelectric focusing planar test according to Embodiment 4 of the present application. [Figure 2] This figure shows the morphological characteristics of the TaC coating layer in crucible 1# according to Example 5 of the present application. [Modes for carrying out the invention]

[0049] The present application will be described in detail below with reference to the examples, but the present application is not limited to these examples.

[0050] Unless otherwise specified, all raw materials used in the embodiments of this application were purchased through commercial channels.

[0051] The stress principle using the Raman test substrate in the following examples is as follows: A Raman spectrometer focuses a monochromatic laser beam of a constant wavelength through a series of optical paths using an objective lens and irradiates it onto the substrate surface. The photons of the laser beam interact with the substrate crystal lattice, causing photon scattering. The occurrence of Raman scattering is related to the crystal lattice vibrations of the substrate itself. If residual stress exists in the sample, tensile or compressive stress causes the atomic bond length to elongate or shorten accordingly, thereby decreasing or increasing the vibration frequency of the atoms. In the Raman spectroscopic diagram, the peak bit of the characteristic peak of the substrate shifts to a lower or higher frequency, i.e., a peak bit shift occurs.

[0052] (Example 1) Quantitative test method for substrate stress First, a silicon carbide substrate is tested using a single-crystal XRD diffractometer, and the substrate is analyzed based on the reciprocal lattice vector space. The parameters of a standard sample 4H-silicon carbide substrate are calculated and compared with the standard parameters. Then, based on the Bragg diffraction equation 2dsinθ=nλ, where d is the pitch of the crystal plane, θ is the diffraction half-angle, n is the diffraction order, and λ is the wavelength, θ and d of different substrates are tested, and it is found that stretching changes corresponding to the pitch d of the crystal plane occur due to tensile and compressive stress inside the crystal. Different crystal plane pitches d0 and d x Based on d x This can be divided into d1 and d2, where d0 is the theoretical pitch of the crystal planes, d1 is the pitch of the crystal planes subjected to tensile stress and d1 is greater than d0, and d2 is the pitch of the crystal planes subjected to compressive stress and d2 is smaller than d0. The pitch difference Δd of the crystal planes at different positions is calculated as Δd = d1 - d0 or Δd = d0 - d2.

[0053] Based on Raman testing, the actual peak bits of the substrate are obtained, and the shift amount Δv of the peak bits at different test points is calculated by comparing it with the standard peak bits. Combining this with Hooke's Law or the stress-strain formula, the stress value σ at different test points of the silicon carbide substrate is calculated. The ratio of the peak bit shift amount to the stress value at different test points is then the ratio of the peak bit shift amount to the stress conversion coefficient μ, and its range is -125 to -2500.

[0054] The formula for Hooke's Law is σ / S = E × (Δd / d0), where S is the area under σ action and is a fixed value. When Hooke's Law is applied to microscopic calculations, S is the microscopic area, which can be written as having a unit of 1. E is the Young's modulus of the substrate, a predetermined parameter of the substrate. Δd is the pitch difference of the crystal planes at different positions, d0 is the theoretical pitch of the crystal planes, and σ is the stress value that needs to be calculated.

[0055] The formula for calculating stress deformation is σ = E × Δd, where E is the Young's modulus of the substrate, a predetermined parameter of the substrate, Δd is the pitch difference of the crystal planes at different locations, and σ is the stress value that needs to be calculated.

[0056] First, Raman test points can be set, allowing for single-point testing, and mapping scans can be performed across the entire substrate. Therefore, single-point testing involves setting fragmentary coordinate points, while mapping scans involve 10x10 to 50x50 or more points. Based on the properties of the substrate, an appropriate test method can be selected to determine the Raman test coordinate diagram for the silicon carbide substrate.

[0057] Next, the results obtained from the test are output, which include, but are not limited to, editable peak bits, peak intensity data, and mapping.

[0058] Subsequently, peak bit fitting is performed on the output data. The fitting function includes, but is not limited to, Gaussian, Lorentz, GaussLor, AGauss, Aloren, and AGaussLor, to obtain accurate peak bit values ​​and peak bit offsets, ensuring that the peak bits have at least one decimal place.

[0059] Finally, the stress value of silicon carbide is calculated based on the peak bit shift amount calculated in (1), the stress conversion coefficient μ, and the peak bit shift amount Δv obtained by the Raman test. The formula is σ(MPa) = μ × Δv(cm -1 ) and finally obtain a stress test image of the substrate.

[0060] Because SiC is subjected to a large amount of external mechanical stress during the wafer manufacturing process, which involves a series of processing steps from the crystalline material, it is difficult to guarantee that the surface is perfectly horizontal; in other words, a certain degree of curvature exists. To eliminate the influence of the curvature of the substrate surface on the substrate surface stress and internal stress tests, it is first necessary to perform a curvature test on the substrate. Based on this curvature, the equipitch focusing test is selected for the substrate surface stress, and the equipitch focusing test or isofocusing plane test is selected for the substrate internal stress.

[0061] After the curvature test, it was necessary to calibrate the tester using single-crystal silicon. First, a Z-axis mapping test was performed on the single-crystal Si to confirm that it was a Gaussian distribution. The surface of the single-crystal Si substrate was set as the zero point, and an axial mapping test was performed from -100 μm within the single-crystal Si substrate to +20 μm above the single-crystal Si substrate. The change in Raman signal intensity was analyzed. Since the single-crystal Si substrate is an opaque material, its peak intensity exhibited a Gaussian distribution. The zero point position on the surface was set as the axis of symmetry, and a surface focusing single-point test was then performed on the single-crystal Si, and the single-crystal Si Raman peak bit was set to 520.70 cm using a correction device. -1 This indicates that the device's optical path has completed calibration.

[0062] Using a corrected tester, an axial test in the Z direction was performed on a 4H-SiC substrate. The Raman laser beam was from -100 μm inside the 4H-SiC substrate to +20 μm above the 4H-SiC substrate. 4H-SiC and other silicon carbides have the characteristics of transparent materials, so the Raman laser beam can be incident into the interior of the silicon carbide substrate, and the feedback signal gradually weakens as the incident depth of the Raman laser beam increases. Other silicon carbides include semi-insulating, conductive, and different crystal types such as 3C, 4H, 15R, and 6H silicon carbide substrates, and therefore the Raman peak intensity of the silicon carbide substrate gradually decreases as the incident depth of the Raman laser beam into the interior of the silicon carbide substrate increases.

[0063] The Raman instrument's automatic focusing function allows it to automatically set to search for the strongest peak of the test signal on the substrate surface, thereby enabling measurement of the curvature of the substrate surface. The specific method is as follows: In S1, first, the horizontal alignment of the test platform is ensured, a predetermined origin (0,0) is selected on the test platform, and this point is denoted as 3D coordinate (0,0,0). In S2, the objective lens is raised to a height where it does not contact the substrate based on the substrate thickness, and one point is automatically focused to the position on the substrate surface where the test signal is strongest, denoted as (a,b,c). Subsequently, the downward H distance of the Raman laser beam is adjusted, and the position is focused to (a,b,cH). In S3, the Raman laser beam test position is adjusted to (0,0,cH) to perform an isofocusing plane test, and a two-dimensional peak intensity scan map of the substrate surface is obtained. Based on the peak intensity distribution relationship of the silicon carbide substrate and the two-dimensional peak intensity scan map of the substrate surface, the distance between the focusing plane of the Raman laser beam and the silicon carbide substrate surface can be obtained, thereby obtaining the curvature within the substrate surface and obtaining a curvature distribution map within the substrate surface.

[0064] The substrate has thermal stress and residual stress inside the crystal due to the influence of the temperature field and the introduced coating during the crystal growth process. The crystal is subjected to a series of cutting, polishing, and grinding processes to obtain the substrate. During the processing of the crystal to form the substrate, surface crystal lattice distortion occurs, thereby generating surface stress. Furthermore, in order to distinguish between the surface stress and the internal stress of the substrate, an axial test, i.e., Z-direction mapping, is performed on the substrate based on the translucency characteristics of silicon carbide. The substrate surface is set as the zero point, and the test range of the axial test is from -150 μm to -30 μm inside the substrate to above the substrate surface.

[0065] In the test, the white spots of the laser beam are gradually focused upward from within the substrate to the damaged stress extension layer, then to the surface damaged layer, and then further upward to the air, determining the thickness of the damaged layer and the damaged stress extension layer on the substrate surface. The substrate surface is set as the zero point, and the test range of the axial test is such that the white spots of the Raman laser beam gradually move upward from -100 μm within the substrate to 30 μm above the substrate surface. Five points are tested at the top, bottom, left, right, and center of the substrate. The five points on the substrate show the same expression trend, with differences in the maximum value of negative stress only at the surface damaged areas.

[0066] The white spots of the laser beam gradually move upward from -100 μm to -30 μm within the substrate. At this point, the stress fluctuates up and down along the horizontal line, indicating a uniform internal stress state. When the laser beam focuses on the damaged stress extension layer, the stress curve begins to show a downward trend at -30 μm, and the downward trend gradually increases. As it approaches the zero point on the substrate surface, it reaches a maximum value of negative stress, which is expressed as the stress value of the surface damaged layer. Finally, the laser beam continues to move upward and gradually detaches from the substrate surface into the air, and the stress is expressed as rapidly increasing in a turbulent fluctuation. Therefore, based on the axial test results of the substrate, the thickness of the surface damaged layer and the damaged stress extension layer of the substrate can be determined, thereby classifying the stress in the substrate into surface stress and internal stress, providing a theoretical basis for subsequent quantitative tests of surface stress and internal stress in the substrate.

[0067] During the crystal growth process of silicon carbide using the PVT method, residual stress is introduced into the crystal due to changes in the potential temperature gradient and intrusion of coatings. This residual stress is distinguished from surface processing stress by its internal location at a certain depth from the surface, where it reacts to the crystal growth thermal stress. Because silicon carbide is subjected to a large amount of external mechanical stress during the process of obtaining a substrate from the crystal, it is difficult to guarantee that the surface is perfectly horizontal; that is, a certain degree of curvature exists on the substrate surface. Example 2 describes a method for quantitatively characterizing the internal stress of silicon carbide, unaffected by surface curvature, by obtaining the curvature of the substrate surface using a Raman test. This example relates to a method for quantitatively characterizing the internal stress of silicon carbide, in conjunction with its physical properties, without being affected by surface curvature.

[0068] Based on the light-transmitting physical properties of silicon carbide, methods for quantitatively characterizing stress within a substrate using Raman spectroscopy are divided into equipitch focusing tests and isofocusing plane tests, specifically as follows.

[0069] 1. Equal-pitch focusing test The principle of the equipitch focusing test is as shown in Figure 1(a). In this test method, the focusing depth (d) of the Raman laser beam is focused into the substrate at different positions, and the transmission depth (D) of the Raman laser beam is determined based on the following formula. p Calculate ).

[0070] D p = 1 / α = λ / 4πk

[0071] In the above equation, D p D is the transmission depth in the sample under Raman excitation, α is the absorption coefficient, λ is the wavelength of the laser light, and k is the attenuation coefficient. Based on the above equation, the transmission depth (D p The specific value of ) is calculated, and an appropriate focusing depth (d) is selected. The requirements that need to be met are the transmission depth (D pSince the focusing depth (d) is greater than the focusing depth (d) and the test was conducted on internal stress, the focusing depth d must be greater than the thickness of the surface damage layer and the damage stress extension layer obtained in Example 3.

[0072] Next, the diameter (D) of the white spot in the Raman laser light is calculated based on the following formula.

[0073] D = 2λ / N a 2 Here, λ is the wavelength of the laser light, and N a This is the numerical hole diameter of the device.

[0074] The diameter of the white spot (D) is the minor axis diameter of the white spot in the laser beam. The larger this value, the stronger the Raman peak intensity at each test position, and the more impurity peaks there are, affecting the accuracy of Raman detection. Once λ is determined, N a If the numerical value is large, the peak intensity will be very weak, which will also affect the accuracy of Raman detection, and therefore the appropriate λ and N values ​​will be necessary. a This requires selecting an option to improve the accuracy of the test method.

[0075] Based on the properties of the silicon carbide material, the peak intensity is strongest when the laser beam is focused on the substrate surface. Therefore, the automatic focusing function of the Raman tester can be used to focus the laser beam on the substrate surface. Subsequently, based on the determined focusing depth (d), equipitch focusing tests can be performed at different positions on the substrate, meaning the distance between the focusing plane and the substrate surface is always equal. This method is unaffected by the curvature or warping of the substrate surface, and the acquired Raman peak intensity is always maintained at the maximum value in the axial direction of the test point. The stress quantitative test method of Example 1 is used to obtain a diagram of the internal stress test results of the substrate.

[0076] 2. Isofocusing Planar Test The principle of the isofocusing plane test is as shown in Figure 1(b). In this test method, the focusing depth (d) changes as the white spots of the Raman laser light are focused into the substrate at different positions. In both cases, the white spots of the laser light are focused into the same plane inside the substrate, and the transmission depth (D) of the Raman laser light is determined based on the same formula as in the equipitch focusing test. p ) is calculated and an appropriate focusing depth (d) is selected using a similar selection method.

[0077] In this test method, since the focusing plane is equal, and in combination with the properties of silicon carbide, the peak intensity is strongest when the laser beam is focused on the substrate surface, and the signal gradually attenuates as the focusing depth increases. The effective signal-to-noise ratio was tested using a Raman tester, and it was discovered that the effective signal-to-noise ratio and the focusing depth (d) of the Raman laser beam are inversely correlated. Therefore, in this test method, the focusing depth (d) at different positions must be matched to the curvature of the substrate. The effective peak bit shift amount is obtained using the stress quantitative test method of Example 1, thereby reducing the error in the internal stress test and obtaining a diagram of the internal stress test results for the substrate.

[0078] Using the two types of equal-pitch focusing and iso-focusing plane test methods described above, the peak bit shift amount inside the substrate was obtained, and the distribution of internal stress in the substrate was calculated. The error in the internal stress values ​​of the substrate calculated using the two methods, equal-pitch focusing test and iso-focusing plane test, was ≤0.2 MPa, and the difference in peak bit values ​​obtained using the two methods, equal-pitch focusing test and iso-focusing plane test, was <0.0005 cm. -1 The two methods described above can verify the accuracy of the quantitative detection method of this application, and the accuracy of the test method can be further improved based on the setting of the focusing depth and the diameter of the white spot.

[0079] (Examples) This embodiment relates to a method for manufacturing a TaC coating layer having through-holes. Slurry preparation: Step (1) involves mixing 25-40 parts TaC powder, 5-8 parts Ta2O5, 20-35 parts solvent, 15-25 parts auxiliary solvent, 1-3 parts air entrainer, 2-5 parts thickener, 5-15 parts linear saturated aliphatic hydrocarbon, and 1-3 parts carbon powder, and stirring at 700-800 r / min for 2-3 hours to obtain a slurry containing TaC. The process includes (2) spraying the slurry onto a graphite substrate, raising the temperature to 60-80°C and maintaining it for 2-4 hours to burst bubbles, then sintering at 180-220°C for 8-12 hours at 700-900 mbar and 500-700°C for 3-5 hours, after which the pressure is increased to 1000-1100 mbar and the temperature is increased to 1600-1800°C for 8-10 hours, after which the pressure is adjusted to 145-155 kPa and the temperature is increased to 2200-2400°C for 14-16 hours, and finally lowering the temperature to room temperature and atmospheric pressure within 15-25 hours, thereby producing a TaC coating layer having through-holes.

[0080] Crucibles 1#-8# and comparative crucibles D1#-D3# were manufactured using the above manufacturing method, specifically as follows: (1) Slurry preparation: Mix 32.7 parts of TaC powder, 6 parts of Ta2O5, 25 parts of ethanol, 20 parts of tetrahydrofuran, 1.5 parts of saponin air entrainer, 3.5 parts of cellulose ether, 10 parts of linear saturated aliphatic hydrocarbon, and 1.3 parts of carbon powder, and stir at 700 r / min for 2.5 hours to obtain a slurry containing TaC. (2) The slurry from step (1) is electrostatically sprayed onto the surface of the graphite substrate, the temperature is raised to 70°C and maintained for 3 hours to burst the bubbles, then dried at 200°C for 10 hours, followed by sintering at 800 mbar and 600°C for 4 hours, then the pressure is increased to 1000 mbar and the temperature to 1700°C for 9 hours, then the pressure is adjusted to 150 kPa and the temperature to 2300°C for 15 hours, and finally the temperature is reduced to room temperature and atmospheric pressure within 20 hours, i.e., crucible #1 is produced.

[0081] (1) Slurry preparation: Mix 25 parts TaC powder, 5 parts Ta2O5, 20 parts ethanol, 15 parts tetrahydrofuran, 1 part saponin air entrainer, 2 parts cellulose ether, 5 parts linear saturated aliphatic hydrocarbon, and 1 part carbon powder, and stir at 800 r / min for 2 hours to obtain a slurry containing TaC. (2) The slurry from step (1) is electrostatically sprayed onto the surface of the graphite substrate, the temperature is raised to 60°C and maintained for 4 hours to burst the bubbles, then dried at 180°C for 12 hours, followed by sintering at 700 mbar and 500°C for 5 hours, then the pressure is increased to 1000 mbar and the temperature is raised to 1600°C for 10 hours, then the pressure is adjusted to 145 kPa and the temperature is raised to 2200°C for 16 hours, and finally the pressure is reduced to room temperature and atmospheric pressure within 15 hours, i.e., crucible #2 is produced.

[0082] (1) Slurry preparation: Mix 40 parts TaC powder, 8 parts Ta2O5, 35 parts ethanol, 25 parts tetrahydrofuran, 3 parts saponin air entrainer, 5 parts cellulose ether, 15 parts linear saturated aliphatic hydrocarbon, and 3 parts carbon powder, and stir at 700 r / min for 3 hours to obtain a slurry containing TaC. (2) The slurry from step (1) is electrostatically sprayed onto the surface of the graphite substrate, the temperature is raised to 80°C and maintained for 2 hours to burst the bubbles, then dried at 220°C for 8 hours, followed by sintering at 900 mbar and 700°C for 3 hours, then the pressure is increased to 1100 mbar and the temperature to 1800°C for 8 hours, then the pressure is adjusted to 155 kPa and the temperature to 2400°C for 14 hours, and finally the mixture is reduced to room temperature and atmospheric pressure within 25 hours, i.e., crucible #3 is produced.

[0083] The difference between this embodiment and crucible 1# is that in step (1), the raw materials are mixed and then stirred at 700 r / min for 5 minutes, while all other conditions are the same as for crucible 1#, thus producing crucible 4#.

[0084] The difference between this embodiment and crucible 1# is that in step (2), the slurry from step (1) is electrostatically sprayed onto the surface of the graphite substrate, the temperature is raised to 200°C and maintained for 13 hours to burst bubbles and dry, and then sintering is performed. All other conditions are the same as for crucible 1#, and crucible 5# is manufactured.

[0085] The difference between this embodiment and crucible 1# is that in step (2), sintering is carried out at 400 mbar and 400°C for 4 hours, then the pressure is increased to 1000 mbar and the temperature to 1700°C for 9 hours, then the pressure is adjusted to 150 kPa and the temperature to 2300°C for 15 hours, and finally the temperature is reduced to room temperature and atmospheric pressure within 20 hours. All other conditions are the same as for crucible 1#, i.e., crucible 6# is manufactured.

[0086] The difference between this embodiment and crucible 1# is that in step (2), sintering is carried out at 800 mbar and 600°C for 4 hours, then the pressure is increased to 1000 mbar and the temperature to 1200°C for 9 hours, then the pressure is adjusted to 150 kPa and the temperature to 2300°C for 15 hours, and finally the temperature is reduced to room temperature and atmospheric pressure within 20 hours. All other conditions are the same as for crucible 1#, i.e., crucible 7# is manufactured.

[0087] The difference between this embodiment and crucible 1# is that in step (2), sintering is carried out at 800 mbar and 600°C for 4 hours, then the pressure is increased to 1000 mbar and the temperature to 1700°C for 9 hours, then the pressure is adjusted to 120 kPa and the temperature to 2500°C for 15 hours, and finally the temperature is reduced to room temperature and atmospheric pressure within 20 hours. All other conditions are the same as for crucible 1#, i.e., crucible 8# is manufactured.

[0088] In this comparative example, the TaC coating layer was manufactured using the CVD method, in contrast to crucible #1, and the details are as follows.

[0089] A graphite substrate is placed in a chemical vapor deposition furnace, the vacuum is reduced to below 50 Pa, and the temperature is then raised to 1200°C. A mixed gas consisting of ethane, TaCl5, H2, and argon gas as a carrier is then flowed through the furnace, maintaining a stable flow rate. The furnace pressure is maintained at 5000 Pa. First, the molar ratio of TaCl5 to ethane gas is controlled to 1.2:1 to form a transition layer with a thickness of 50 μm. Subsequently, the molar ratio of TaCl5 to ethane gas is controlled to 3:1 to form a TaC coating layer with a thickness of 60 μm, thereby producing comparative crucible D1#.

[0090] This comparative example differs from crucible 1# in that the raw materials in the slurry of step (1) are different, with BYK defoaming agent used instead of saponin-based air-entraining agent, and all other conditions are the same as crucible 1#, i.e., comparative crucible D2# is produced.

[0091] This comparative example differs from crucible 1# in that the raw materials in the slurry of step (1) are different, using TaCl5 instead of Ta2O5, while all other conditions are the same as crucible 1#, i.e., comparative crucible D3# is produced.

[0092] Cross-sectional SEM tests and EDS energy spectral analyses were performed on the manufactured crucibles 1#-8# and comparative crucibles D1#-D3#. As can be seen in Figure 2, Figure 2(a) shows the microscopic morphology of the manufactured TaC coated layer with through-holes, where holes of similar morphology and size are uniformly distributed on its surface. Figure 2(b) shows the coated layer after polishing until the underlying graphite is exposed. Holes are still visible at this stage, thus proving that the formed holes are through-holes. XRD tests were performed on the TaC coated layer, and the test results show that the TaC crystal grains in the TaC coated layer grow oriented along the {200} and {220} directions. The thickness of the formed TaC coated layer, the average pore diameter of the through-holes in the TaC coated layer, the density, and the size of the TaC crystal grains were tested, and the test results are shown in Table 1.

[0093] [Table 1]

[0094] High-temperature erosion and thermochemical erosion experiments were performed on the manufactured crucibles 1#-8# and comparative crucibles D1#-D3# to simulate the growth environment used by the PVT method for third-generation semiconductor silicon carbide crystals. Crucibles 1#-8# and comparative crucibles D1#-D3# were subjected to silicon carbide atmosphere erosion at 2300°C and maintained for 100 hours. The test results are shown in Table 2.

[0095] [Table 2]

[0096] Table 2 shows that the erosion weight loss rate = {(weight before experiment - weight after experiment) / weight before experiment} × 100%. The above high-temperature erosion and thermochemical erosion experiments demonstrate that the TaC coating layer of this embodiment can improve the protective performance against the graphite substrate, extend the durability and corrosion resistance of the TaC coating layer, and thereby extend the service life of the crucible.

[0097] This embodiment relates to a method for manufacturing a silicon carbide single crystal with a diameter of 150 mm. Step (1) involves adding silicon carbide powder to the crucible, 10 inside the reactor -6 The vacuum is reduced to below mbar, followed by the flow of high-purity inert gas up to 300-500 mbar. This process is repeated 2-3 times, and finally the furnace body is filled to 10 -6 Step (2) involves vacuuming to below mbar, Step (3) involves flowing high-purity inert gas into the furnace body, raising the pressure to 10-100 mbar within 1-3 hours, and maintaining the pressure by continuously flowing high-purity inert gas. In the crystal growth stage, if the pressure is kept constant, the temperature inside the furnace is raised to the single crystal growth temperature of 2200K to 2800K within 3 to 5 hours, and the growth time is 30 to 150 hours (step 4), The process includes step (5) obtaining a silicon carbide single crystal after the growth of the single crystal is complete, opening the furnace body, and removing the graphite crucible.

[0098] Equal amounts of silicon carbide powder were added to crucibles 1#-3# and comparative crucibles D1#-D3#, respectively, and the furnace body was filled to 10 -6 The furnace body is evacuated to below mbar, then high-purity inert gas is flowed up to 300 mbar, and this process is repeated three times, until finally the furnace body reaches 10 mbar. -6 The furnace is evacuated to below mbar, high-purity inert gas is flowed into the furnace, the pressure is raised to 30 mbar within 1 hour, and the pressure is maintained by continuously flowing high-purity inert gas. During the crystal growth stage, if the pressure is maintained without changing, the temperature inside the furnace is raised to the single crystal growth temperature of 2500 K within 3 hours, the growth time is 70 hours, the single crystal growth is completed, the furnace is opened, the graphite crucible is removed, and silicon carbide single crystals 1#-3# and comparative silicon carbide single crystals D1#-D3# are obtained.

[0099] After performing the same cutting, polishing, mechanical polishing, and chemical polishing on the silicon carbide single crystals 1#-3# and comparative silicon carbide single crystals D1#-D3#, silicon carbide substrates 1#-3# and comparative silicon carbide substrates D1#-D3# were obtained, and internal stress tests were performed using the methods of Examples 1-4, with the results shown in Tables 3 and 4.

[0100] [Table 3]

[0101] [Table 4]

[0102] Planar pattern detection was performed on the silicon carbide substrates 1#-3#. The curvature of all silicon carbide substrates 1#-3# was within the range of -50 to 50 μm, the local thickness variation in the central region was within the range of 0.1 to 2 μm, and the local thickness variation in the annular region was within the range of 0.1 to 2 μm.

[0103] The stress values ​​in Tables 3 and 4 are all relative stress values. The sign of the stress value in this application indicates whether the substrate is subjected to tensile or compressive stress. Currently, when evaluating and explaining the final stress of a single substrate, it is always obtained by subtracting the minimum value from the maximum value of the above stress values. For example, if the axial stress of the silicon carbide substrate 1# square region is -0.5 to 0.7 MPa, in this field, when roughly representing the axial stress of the silicon carbide substrate 1# square region, it can also be explained as having a stress value of 1.2 MPa.

[0104] The data in Tables 3 and 4 above are based on the case where silicon carbide substrate sheets are manufactured using only substrates produced in crucibles 1#-3# and comparative crucibles D1#-D3# manufactured in Example 5, and using the same crystal growth and processing method as in Example 6, with crucibles 4#-8# used. When the parameters in Tables 3 and 4 are characterized using absolute stress, the width of the annular region extending inward from the edge of the substrate is 5 to 30 mm, the radial internal stress of the square region is -10 to 10 MPa, the radial internal stress of the annular region is 5 to 15 MPa, and 1.8 MPa ≤ S max2 -S max1 The requirements can be met such that the pressure is ≤5.2MPa, -5MPa≦S1-S2≦5MPa, 0.3≦S1 / S3≦0.94, 0MPa≦△S1≦10MPa, 0MPa≦△S2≦5MPa, the internal stress in the axial direction of the square region is -10 to 10MPa, the internal stress in the axial direction of the annular region is -15 to 15MPa, 0MPa≦△S3≦10MPa, 0MPa≦△S4≦15MPa, -5MPa≦S4-S5≦5MPa, and 0.1≦S4 / S6≦0.8.

[0105] The above descriptions are merely embodiments of the present application, and the scope of protection of the present application is not limited to these specific embodiments but is determined by the claims of the present application. To those skilled in the art, the present application is subject to various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the technical idea and principles of the present application should be included within the scope of protection of the present application.

Claims

1. A silicon carbide substrate of high crystal quality, wherein the silicon carbide substrate has a diameter of 150 mm or more, and the silicon carbide substrate includes a first main surface and a second main surface. The first main surface has a central region and an annular region surrounding the central region, and the annular region extends inward from the edge of the substrate and has a width of 5 to 30 mm. A silicon carbide substrate with high crystal quality, characterized in that the central region is divided into square regions, each having a side length of 5 mm, the internal stress in each square region is smaller than the internal stress in the annular region, and the internal stress is the stress value detected at a location extending at least 30 μm perpendicularly into the silicon carbide substrate from the first or second main surface.

2. The silicon carbide substrate according to claim 1, characterized in that, in either the plane parallel to the first main surface and / or the second main surface, the radial internal stress in the square region is -10 to 10 MPa, and the radial internal stress in the annular region is 5 to 15 MPa.

3. In either of the planes parallel to the first principal plane and / or the second principal plane, S ｍａｘ１ This represents the maximum absolute value of the radial internal stress in all of the aforementioned square regions, S ｍａｘ２ This represents the maximum absolute value of the radial internal stress within the annular region, where 1.8 MPa ≤ S ｍａｘ２ -S ｍａｘ１ The silicon carbide substrate according to claim 1, characterized in that the pressure is ≤ 5.2 MPa.

4. The silicon carbide substrate according to claim 1, characterized in that, in any plane parallel to the first principal surface and / or the second principal surface, S1 represents the average value of the absolute values ​​of the radial internal stress in all the square regions, S2 represents the average value of the absolute values ​​of the radial internal stress in any square region, and -5 MPa ≤ S1 - S2 ≤ 5 MPa.

5. The silicon carbide substrate according to claim 1, characterized in that, in either the first principal surface and / or the plane parallel to the second principal surface, S1 represents the average value of the absolute values ​​of the radial internal stress in all the square regions, S3 represents the average value of the absolute values ​​of the radial internal stress in the annular region, and 0.3 ≤ S1 / S3 ≤ 0.

94.

6. In any plane parallel to the first major surface and / or the second major surface, S ｍａｘ３ represents the maximum value of the in-plane stress in the radial direction in any of the square regions, and S ｍｉｎ３ represents the minimum value of the radial stress in any of the square regions, and ΔS １ = S ｍａｘ３ - S ｍｉｎ３ and 0 MPa ≤ ΔS １ ≤ 10 MPa. In either of the planes parallel to the first principal plane and / or the second principal plane, S ｍａｘ４ This represents the maximum value of the internal stress in the radial direction in the annular region, S ｍｉｎ４ This represents the minimum value of radial stress in the annular region, and ΔS ２ = S ｍａｘ４ -S ｍｉｎ４ Therefore, 0 MPa ≤ ΔS ２ The silicon carbide substrate according to claim 1, characterized in that the pressure is ≤ 5 MPa.

7. The silicon carbide substrate according to claim 1, characterized in that an axial test is performed extending perpendicularly into the silicon carbide substrate from any point on the first main surface, and the axial internal stress in the square region is -10 to 10 MPa, and the axial internal stress in the annular region is -15 to 15 MPa.

8. An axial test is performed by extending perpendicularly from any point on the first main surface into the silicon carbide substrate, S ｍａｘ５ Let S be denoted as the maximum value of the axial stress along the axis of the square region, ｍｉｎ５ Let ΔS be denoted as the minimum value of the axial stress along the axis of the square region, ３ = S ｍａｘ５ -S ｍｉｎ５ Therefore, 0 MPa ≤ ΔS ３ The pressure is ≤ 10 MPa, An axial test is performed by extending perpendicularly from any point on the first main surface into the silicon carbide substrate, S ｍａｘ６ Let S be denoted as the maximum axial stress in the annular region. ｍｉｎ６ Let ΔS be denoted as the minimum value of axial stress in the annular region. ４ = S ｍａｘ６ -S ｍｉｎ６ Therefore, 0 MPa ≤ ΔS ４ The silicon carbide substrate according to claim 7, characterized in that the pressure is ≤ 15 MPa.

9. The silicon carbide substrate according to claim 7, characterized in that S4 represents the average value of the absolute values ​​of the internal stress in the axial direction in all of the square regions, S5 represents the average value of the absolute values ​​of the internal stress in the radial direction in any of the square regions, and -5 MPa ≤ S4 - S5 ≤ 5 MPa.

10. The silicon carbide substrate according to claim 1, characterized in that S4 represents the average value of the absolute values ​​of the axial internal stress in all the square regions, and S6 represents the average value of the absolute values ​​of the axial internal stress in the annular region, and 0.1 ≤ S4 / S6 ≤ 0.8.

Images