Silicon carbide substrate and method for manufacturing silicon carbide substrate

Abstract

The ratio of the number of pits divided by the number of screw dislocations is 1% or less. The surface roughness of a first main surface is 0.15 nm or less. The absolute value of the difference between a first wave number and a second wave number is 0.2 cm-1 or less, and the absolute value of the difference between a first full width at half maximum and a second full width at half maximum is 0.25 cm-1 or less, where the first wave number is the average value of wave numbers indicating the peak of the Raman spectrum of silicon carbide in a first square region, the second wave number is the average value of wave numbers indicating the peak of the Raman spectrum of silicon carbide in a second square region, the first full width at half maximum is the average value of the full width at half maximum of the peak of the Raman spectrum of silicon carbide in the first square region, and the second full width at half maximum is the average value of the full width at half maximum of the peak of the Raman spectrum of silicon carbide in the second square region.

Description

Technical Field

[0001] The present disclosure relates to a silicon carbide substrate and a method for manufacturing a silicon carbide substrate. This application claims priority based on Japanese Patent Application No. 2019-218125, filed on December 2, 2019. All the descriptions described in the Japanese patent application are incorporated herein by reference.

Background Art

[0002] Japanese Patent Application Laid-Open No. 2014-210690 (Patent Document 1) describes that chemical mechanical polishing is performed on a silicon carbide single crystal substrate.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

[0004] The silicon carbide substrate according to the present disclosure includes a first main surface and a second main surface on the opposite side of the first main surface. The silicon carbide substrate includes screw dislocations and pits having a maximum diameter in a direction parallel to the first main surface of 1 μm or more and 10 μm or less. When observing screw dislocations and pits on the first main surface, the ratio of the number of pits divided by the number of screw dislocations is 1% or less. The surface roughness of the first main surface is 0.15 nm or less. In a first square region including screw dislocations and having a side length of 200 μm, the average value of the wave numbers indicating peaks corresponding to the folded mode of the longitudinal optical branch of the Raman spectrum of silicon carbide is defined as the first wave number. In a second square region not including screw dislocations and having a side length of 200 μm, the average value of the wave numbers indicating peaks corresponding to the folded mode of the longitudinal optical branch of the Raman spectrum of silicon carbide is defined as the second wave number. In the first square region, the average value of the full width at half maximum of the peaks corresponding to the folded mode of the longitudinal optical branch of the Raman spectrum of silicon carbide is defined as the first full width at half maximum. And in the second square region, when the average value of the full width at half maximum of the peaks corresponding to the folded mode of the longitudinal optical branch of the Raman spectrum of silicon carbide is defined as the second full width at half maximum, the absolute value of the difference between the first wave number and the second wave number is 0.2 cm -1 or less, and the absolute value of the difference between the first full width at half maximum and the second full width at half maximum is 0.25 cm -1 or less.

[0005] The manufacturing method of the silicon carbide substrate according to the present disclosure includes the following steps. A silicon carbide single crystal substrate having a first main surface and a second main surface on the opposite side of the first main surface is prepared. On the first main surface, the silicon carbide single crystal substrate is mechanically polished. After the step of mechanically polishing the silicon carbide single crystal substrate, the silicon carbide single crystal substrate is etched. After the step of etching the silicon carbide single crystal substrate, on the first main surface, the silicon carbide single crystal substrate is chemically mechanically polished using abrasive grains and an oxidizing agent. In the step of mechanically polishing the silicon carbide single crystal substrate, a damaged layer is formed on the first main surface. In the step of etching the silicon carbide single crystal substrate, the damaged layer is removed. In the step of chemically mechanically polishing the silicon carbide single crystal substrate, when the relationship between the surface roughness of the first main surface as the vertical axis and the concentration of the oxidizing agent as the horizontal axis is approximated by a first quadratic curve, the concentration of the oxidizing agent is within a range where the surface roughness is 1.5 times or less the minimum value of the first quadratic curve, and the polishing rate of the silicon carbide single crystal substrate is 0.2 μm / hour or more.

Brief Description of the Drawings

[0006]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

[0007] [Problems to be Solved by the Present Disclosure] The object of the present disclosure is to suppress the formation of pits after epitaxial growth. [Effect of the present disclosure] According to the present disclosure, it is possible to provide a silicon carbide substrate capable of suppressing the formation of pits after epitaxial growth and a method for manufacturing the silicon carbide substrate. [Description of embodiments of the present disclosure] (1) The silicon carbide substrate 10 according to the present disclosure includes a first main surface 1 and a second main surface 2 on the opposite side of the first main surface 1. The silicon carbide substrate 10 includes a screw dislocation 13 and a pit 11 having a maximum diameter in a direction parallel to the first main surface 1 of 1 μm or more and 10 μm or less. When observing the screw dislocation 13 and the pit 11 on the first main surface 1, the ratio obtained by dividing the number of pits 11 by the number of screw dislocations 13 is 1% or less. The surface roughness of the first main surface 1 is 0.15 nm or less. In a first square region 14 including the screw dislocation 13 and having a side length of 200 μm, the average value of the wave numbers indicating the peaks corresponding to the folded mode of the longitudinal optical branch of the Raman spectrum of silicon carbide is defined as the first wave number. In a second square region 15 not including the screw dislocation 13 and having a side length of 200 μm, the average value of the wave numbers indicating the peaks corresponding to the folded mode of the longitudinal optical branch of the Raman spectrum of silicon carbide is defined as the second wave number. In the first square region 14, the average value of the half-widths at half-maximum of the peaks corresponding to the folded mode of the longitudinal optical branch of the Raman spectrum of silicon carbide is defined as the first half-width, and in the second square region 15, the average value of the half-widths at half-maximum of the peaks corresponding to the folded mode of the longitudinal optical branch of the Raman spectrum of silicon carbide is defined as the second half-width. In this case, the absolute value of the difference between the first wave number and the second wave number is 0.2 cm -1 or less, and the absolute value of the difference between the first half-width and the second half-width is 0.25 cm -1 or less.

[0008] (2) According to the silicon carbide substrate 10 according to (1) above, the ratio obtained by dividing the number of pits 11 by the number of screw dislocations 13 may be 0.5% or less.

[0009] (3) According to the silicon carbide substrate 10 according to (1) above, the ratio obtained by dividing the number of pits 11 by the number of screw dislocations 13 may be 0.4% or less.

[0010] (4) According to the silicon carbide substrate 10 according to any one of (1) to (3) above, the surface roughness of the first main surface 1 may be 0.1 nm or less.

[0011] (5) According to the silicon carbide substrate 10 according to any one of (1) to (4) above, the diameter of the first main surface 1 may be 150 mm or more.

[0012] (6) According to the silicon carbide substrate 10 according to any one of (1) to (5) above, the surface density of the screw dislocations 13 on the first main surface 1 is 100 cm -2 or more and 5000 cm -2 or less.

[0013] (7) The method for manufacturing the silicon carbide substrate 10 according to the present disclosure includes the following steps. A silicon carbide single crystal substrate 100 having a first main surface 1 and a second main surface 2 on the opposite side of the first main surface 1 is prepared. On the first main surface 1, the silicon carbide single crystal substrate 100 is mechanically polished. After the step of mechanically polishing the silicon carbide single crystal substrate 100, the silicon carbide single crystal substrate 100 is etched. After the step of etching the silicon carbide single crystal substrate 100, on the first main surface 1, the silicon carbide single crystal substrate 100 is chemically mechanically polished using abrasive grains and an oxidizing agent. In the step of mechanically polishing the silicon carbide single crystal substrate 100, a damage layer 23 is formed on the first main surface 1. In the step of etching the silicon carbide single crystal substrate 100, the damage layer 23 is removed. In the step of chemically mechanically polishing the silicon carbide single crystal substrate 100, when the relationship between the surface roughness and the concentration of the oxidizing agent is approximated by a first quadratic curve with the surface roughness of the first main surface 1 as the vertical axis and the concentration of the oxidizing agent as the horizontal axis, the concentration of the oxidizing agent is within a range where the surface roughness is 1.5 times or less the minimum value of the first quadratic curve, and the polishing rate of the silicon carbide single crystal substrate 100 is 0.2 μm / hour or more.

[0014] (8) According to the method for manufacturing the silicon carbide substrate 10 according to (7) above, in the step of chemically mechanically polishing the single crystal silicon carbide substrate 100, when the surface roughness of the first main surface 1 is taken as the vertical axis and the diameter of the abrasive grains is taken as the horizontal axis, and the relationship between the surface roughness and the diameter of the abrasive grains is approximated by a second quadratic curve, the diameter of the abrasive grains may be within a range where the surface roughness is 1.5 times or less the minimum value of the second quadratic curve.

[0015] (9) According to the method for manufacturing the silicon carbide substrate 10 according to (7) or (8) above, the step of etching the single crystal silicon carbide substrate 100 may be performed at 400 °C or lower.

[0016] (10) According to the method for manufacturing the silicon carbide substrate 10 according to any one of (7) to (9) above, the minimum value of the first quadratic curve may be 0.15 nm or less.

[0017] (11) According to the method for manufacturing the silicon carbide substrate 10 according to any one of (7) to (10) above, the abrasive grains may be colloidal silica.

[0018] (12) According to the method for manufacturing the silicon carbide substrate 10 according to any one of (7) to (11) above, the step of etching the single crystal silicon carbide substrate 100 may be performed by immersing the damaged layer 23 in a solution.

[0019] (13) According to the method for manufacturing the silicon carbide substrate 10 according to (12) above, the solution may contain potassium permanganate and potassium hydroxide. [Details of Embodiments of the Present Disclosure] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. In the following drawings, the same or corresponding parts are denoted by the same reference numerals, and the description thereof will not be repeated. In the crystallographic descriptions in this specification, individual orientations are indicated by [], collective orientations by <>, individual planes by (), and collective planes by {}. Also, for negative indices, in crystallography, a "-" (bar) is attached above the number, but in this specification, a negative sign is attached before the number.

[0020] First, the configuration of the silicon carbide substrate according to this embodiment will be described. FIG. 1 is a schematic plan view showing the configuration of the silicon carbide substrate according to this embodiment. FIG. 2 is a schematic cross-sectional view taken along line II-II of FIG. 1.

[0021] As shown in FIGS. 1 and 2, the silicon carbide substrate 10 according to this embodiment mainly has a first main surface 1, a second main surface 2, and an outer peripheral surface 5. As shown in FIG. 2, the second main surface 2 is on the opposite side of the first main surface 1. The silicon carbide substrate 10 is composed of polytype 4H silicon carbide. The silicon carbide substrate 10 contains an n-type impurity such as nitrogen (N). The conductivity type of the silicon carbide substrate 10 is, for example, n-type. The concentration of the n-type impurity in the silicon carbide substrate 10 is, for example, 1×10 17 cm -3 or more and 1×10 20 cm -3 or less.

[0022] As shown in FIG. 1, the maximum diameter A of the first main surface 1 is, for example, 150 mm or more (6 inches or more). The maximum diameter A of the first main surface 1 may be, for example, 200 mm or more (8 inches or more). In this specification, 2 inches means 50 mm or 50.8 mm (25.4 mm / inch × 2 inch). 3 inches means 75 mm or 76.2 mm (25.4 mm / inch × 3 inch). 4 inches means 100 mm or 101.6 mm (25.4 mm / inch × 4 inch). 5 inches means 125 mm or 127.0 mm (25.4 mm / inch × 5 inch). 6 inches means 150 mm or 152.4 mm (25.4 mm / inch × 6 inch). 8 inches means 200 mm or 203.2 mm (25.4 mm / inch × 8 inch).

[0023] The first major surface 1 is, for example, a surface inclined at an off-angle greater than 0° and equal to or less than 8° with respect to the {0001} plane or the {0001} plane. The off-angle may be, for example, 1° or more, or may be 2° or more. The off-angle may be 7° or less, or may be 6° or less. Specifically, the first major surface 1 may be a surface inclined at an off-angle greater than 0° and equal to or less than 8° with respect to the (0001) plane or the (0001) plane. The first major surface 1 may be a surface inclined at an off-angle greater than 0° and equal to or less than 8° with respect to the (000-1) plane or the (000-1) plane. The inclination direction of the first major surface 1 is, for example, the <11-20> direction.

[0024] As shown in FIG. 1, the outer peripheral surface 5 may have, for example, a first flat portion 3 and an arcuate portion 4. The first flat portion 3 extends, for example, along the first direction 101. The arcuate portion 4 is continuous with the first flat portion 3. The outer peripheral surface 5 may have a second flat portion (not shown) extending along the second direction 102. The second direction 102 is, for example, the <1-100> direction. The first direction 101 is parallel to the first major surface 1 and perpendicular to the second direction 102. The first direction 101 is, for example, the <11-20> direction.

[0025] The first major surface 1 is, for example, an epitaxial layer formation surface. From another perspective, a silicon carbide epitaxial layer (not shown) is provided on the first major surface 1. The second major surface 2 is, for example, a drain electrode formation surface. From another perspective, a drain electrode (not shown) of a MOSFET (Metal Oxide Semiconductor Field Effect Transistor) is formed on the second major surface 2.

[0026] As shown in FIG. 2, the silicon carbide substrate 10 includes a plurality of screw dislocations 13, pits 11, and a silicon carbide region 22. The plurality of screw dislocations 13 have a first screw dislocation 6 connected to the pit 11 and a second screw dislocation 7 not connected to the pit 11. From another perspective, the pit 11 is caused by the first screw dislocation 6. The pit 11 opens to the first main surface 1 and does not open to the second main surface 2. The first screw dislocation 6 is connected to the second main surface 2. The second screw dislocation 7 is connected to each of the first main surface 1 and the second main surface 2. From another perspective, the second screw dislocation 7 penetrates the silicon carbide region 22 from the first main surface 1 to the second main surface 2.

[0027] FIG. 3 is an enlarged schematic view of region III in FIG. 2. As shown in FIG. 3, the width (diameter) of the pit 11 decreases as it goes from the first main surface 1 toward the second main surface 2. The pit 11 may be, for example, substantially conical. When viewed in a direction perpendicular to the first main surface 1, the pit 11 is substantially circular. The maximum diameter (first diameter W) of the pit 11 in a direction parallel to the first main surface 1 is 1 μm or more and 10 μm or less. The first diameter W may be 2 μm or more, or may be 3 μm or more. The maximum depth (first depth D) of the pit 11 in a direction perpendicular to the first main surface 1 may be, for example, 3 nm or more and 1 μm or less.

[0028] As shown in FIG. 2, the number of screw dislocations 13 is larger than the number of pits 11. Specifically, when observing the screw dislocations 13 and the pits 11 on the first main surface 1, the ratio of the number of pits 11 divided by the number of screw dislocations 13 is 1% or less. The ratio of the number of pits 11 divided by the number of screw dislocations 13 may be, for example, 0.5% or less, or may be 0.4% or less, or may be 0.3% or less. The lower limit of the ratio of the number of pits 11 divided by the number of screw dislocations 13 is not particularly limited, but may be, for example, 0.01% or more, or may be 0.1% or more.

[0029] The areal density of the screw dislocations 13 on the first main surface 1 is, for example, 100 cm -2 or more and 5000 cm -2The following is the case. The lower limit of the surface density of the screw dislocation 13 on the first major surface 1 is not particularly limited, but for example, it may be 200 cm -2 or more, or may be 500 cm -2 or more. The upper limit of the surface density of the screw dislocation 13 on the first major surface 1 is not particularly limited, but for example, it may be 4500 cm -2 or less, or may be 4000 cm -2 or less.

[0030] (Method for measuring screw dislocation) The number of screw dislocations 13 can be measured, for example, using the X-ray topography method. The measuring device is, for example, the XRTmicron manufactured by Rigaku. Specifically, based on the X-ray topographic image of the first major surface 1 of the silicon carbide substrate 10, the number of screw dislocations 13 can be measured. The X-ray topographic image is obtained by (0008) reflection. A Cu target is used as the X-ray source during measurement. The pixel size of the X-ray camera is 5.4 μm.

[0031] (Method for measuring pit) The number of pits 11 can be measured, for example, using a defect inspection device equipped with a confocal differential interference microscope. The defect inspection device is, for example, the WASAVI series "SICA 6X" manufactured by Lasertec Corporation. The magnification of the objective lens is, for example, 10 times. Specifically, light with a wavelength of 546 nm is irradiated from a light source such as a mercury xenon lamp onto the first major surface 1 of the silicon carbide substrate 10, and the reflected light of the light is observed by a light receiving element such as a CCD (Charge-Coupled Device).

[0032] The difference between the brightness of a certain pixel in the observed image and the brightness of the pixels around the certain pixel is digitized. The detection sensitivity threshold of the defect inspection device is determined using a standard sample. By using the defect inspection device, the diameter of the pits 11 formed in the sample to be measured can be quantitatively evaluated. By observing the first major surface 1 of the silicon carbide substrate 10, pits 11 with a maximum diameter (first diameter W) of 1 μm or more and 10 μm or less are detected.

[0033] The surface roughness of the first main surface 1 is 0.15 nm or less. The surface roughness of the first main surface 1 may be, for example, 0.13 nm or less, or may be 0.11 nm or less. The lower limit of the surface roughness of the first main surface 1 is not particularly limited, but may be, for example, 0.01 nm or more. The surface roughness of the first main surface 1 is defined as the arithmetic mean roughness (Sa). The arithmetic mean roughness (Sa) is a parameter obtained by extending the two-dimensional arithmetic mean roughness (Ra) to three dimensions.

[0034] (Method for measuring surface roughness) The arithmetic mean roughness (Sa) can be measured, for example, by a white light interference microscope. Specifically, the first main surface 1 of the silicon carbide substrate 10 is observed by a white light interference microscope. As the white light interference microscope, for example, BW-D507 manufactured by Nikon Corporation can be used. The measurement range of the arithmetic mean roughness (Sa) is, for example, a square region of 255 μm × 255 μm. The center of the diagonal of the square region is, for example, the center of the first main surface 1. The center of the first main surface 1 is, for example, the center of the circle including the arc-shaped portion 4. One side of the square region is parallel to the first direction 101.

[0035] FIG. 4 is an enlarged schematic view of region IV in FIG. 1. As shown in FIG. 4, the first main surface 1 has a first square region 14 and a second square region 15. The first square region 14 includes a screw dislocation 13. There is a silicon carbide region 22 around the screw dislocation 13. As shown in FIG. 4, the first square region 14 includes the screw dislocation 13 and the silicon carbide region 22. The length of one side of the first square region 14 is 200 μm. That is, the first square region 14 is a square region of 200 μm × 200 μm. The screw dislocation 13 is located at the center of the square. One side of the first square region 14 is parallel to the first direction 101.

[0036] As shown in FIG. 4, the second square region 15 does not contain a screw dislocation 13. The second square region 15 includes a silicon carbide region 22. The length of one side of the second square region 15 is 200 μm. That is, the second square region 15 is a square region of 200 μm × 200 μm. One side of the second square region 15 is parallel to the first direction 101. Each of the first square region 14 and the second square region 15 has Raman characteristics described later.

[0037] First, the configuration of a Raman spectrometer for measuring a Raman spectrum will be described. FIG. 5 is a schematic diagram showing the configuration of the Raman spectrometer.

[0038] As shown in FIG. 5, the Raman spectrometer 30 mainly includes, for example, a light source 32, an objective lens 31, a spectroscope 33, a stage 34, a beam splitter 35, and a detector 38. As the Raman spectrometer 30, for example, LabRAM HR-800 manufactured by HORIBA JOBIN YVON can be used. The light source 32 is, for example, a YAG (Yttrium Aluminum Garnet) laser. The excitation wavelength of the light source 32 is, for example, 532 nm. The laser irradiation intensity is, for example, 10 mW. The measurement method is, for example, backscattering measurement. The magnification of the objective lens 31 is 100 times. The diameter of the measurement region is, for example, 1 μm. The irradiation time of the laser is, for example, 20 seconds. The number of integration times is, for example, 5 times. The grating is 2400 gr / mm.

[0039] Next, a method for measuring a Raman spectrum will be described. First, incident light 36 is emitted from the YAG laser of the light source 32. As shown by the arrow 61 in FIG. 5, the incident light 36 is reflected by the beam splitter 35 and is incident toward the first main surface 1 of the silicon carbide substrate 10. The Raman spectrometer 30 employs, for example, a confocal optical system. In the confocal optical system, a confocal aperture (not shown) having a circular aperture is arranged at a position conjugate to the focal point of the objective lens 31. Thereby, only the light at the focused position can be detected.

[0040] As shown by the arrow 62 in FIG. 5, the Raman scattered light scattered by the silicon carbide substrate 10 passes through the beam splitter 35 and is introduced into the spectroscope 33. In the spectroscope 33, the Raman scattered light is decomposed for each wave number. The Raman scattered light decomposed for each wave number is detected by the detector 38. Thereby, a Raman spectrum is obtained with the wave number on the horizontal axis and the intensity of the Raman scattered light on the vertical axis. The stage 34 can move in a direction parallel to the first main surface 1 of the silicon carbide substrate 10 (the direction of the arrow 63).

[0041] FIG. 6 is a schematic diagram showing the measurement points of the Raman spectrum in the first square region 14. As shown in FIG. 6, the Raman spectrum is measured at a plurality of measurement points in the first square region 14. The measurement points of the Raman spectrum are circular regions with a diameter of about 1 μm indicated by white circles. For example, first, the Raman spectrum is measured at the position of the lower left corner (the first position) of the first square region 14. Next, the stage 34 is moved in a direction parallel to the first main surface 1, and the position of the focus of the incident light 36 is adjusted, for example, upward. Thereby, the Raman spectrum at the second position 20 μm away from the first position in the second direction 102 is measured. As described above, by moving the stage 34 along the direction of the arrow 63, the Raman spectrum is measured at a plurality of measurement points in the first square region 14. The pitch of the measurement positions is, for example, 20 μm. The number of measurement positions is, for example, 10 (the first direction 101) × 10 (the second direction 102) = 100.

[0042] FIG. 7 is a schematic diagram showing measurement points of the Raman spectrum in the second square region 15. As shown in FIG. 7, the Raman spectrum is measured at a plurality of measurement points in the second square region 15. The measurement points of the Raman spectrum are circular regions with a diameter of about 1 μm indicated by white circles. For example, first, the Raman spectrum is measured at the position of the lower left corner (the third position) of the second square region 15. Next, the stage 34 is moved in a direction parallel to the first main surface 1, and the position of the focus of the incident light 36 is adjusted, for example, upward. Thereby, the Raman spectrum at the fourth position 20 μm away from the third position in the second direction 102 is measured. As described above, by moving the stage 34 along the direction of the arrow 63, the Raman spectrum is measured at a plurality of measurement points in the second square region 15. The pitch of the measurement positions is, for example, 20 μm. The number of measurement positions is, for example, 10 (in the first direction 101) × 10 (in the second direction 102) = 100.

[0043] FIG. 8 is a schematic diagram showing an example of the Raman spectrum of the silicon carbide substrate 10. The horizontal axis in FIG. 8 is the wave number (Raman shift). The vertical axis in FIG. 8 is the intensity of the Raman scattered light (Raman intensity). The wavelength of the excitation light of the light source 32 is 514.5 nm. The Raman shift is the difference between the wave number of the excitation light and the wave number of the Raman scattered light of the object to be measured. When the object to be measured is polytype 4H silicon carbide, mainly four peaks are observed in the Raman spectrum. The first peak 41 is Raman scattered light caused by the folded mode of the longitudinal optical (LO) branch. The first peak 41 appears, for example, near 964 cm -1 . The second peak 42 is Raman scattered light caused by the folded mode of the transverse optical (TO) branch. The second peak 42 appears, for example, near 776 cm -1 . The third peak 43 is Raman scattered light caused by the folded mode of the longitudinal acoustic (LA) branch. The third peak 43 appears, for example, near 610 cm -1 . The fourth peak 44 is Raman scattered light caused by the folded mode of the transverse acoustic (TA) branch. The fourth peak 44 appears, for example, near 196 cm -1 .

[0044] FIG. 9 is a schematic diagram showing a Raman spectrum measured in the first square region 14 and a Raman spectrum measured in the second square region 15. The Raman spectrum (first Raman spectrum 51) shown by the solid line in FIG. 9 shows the Raman spectrum of silicon carbide measured in the first square region 14. Using the first Raman spectrum 51, the wave number ν1 of the peak corresponding to the folded mode of the longitudinal optical branch is obtained. Note that the peak corresponding to the folded mode of the longitudinal optical branch is a peak of the Raman spectrum generated due to the folded mode of the longitudinal optical branch. Similarly, using the first Raman spectrum 51, the full width at half maximum Δ1 of the peak corresponding to the folded mode of the longitudinal optical branch is obtained.

[0045] Specifically, the full width at half maximum Δ1 is the full width at half maximum (FWHM). Each of the wave number ν1 and the full width at half maximum Δ1 is obtained at 100 measurement positions in the first square region 14. In the first square region 14, the average value of the wave number ν1 is the first wave number. In the first square region 14, the average value of the full width at half maximum Δ1 is the first full width at half maximum.

[0046] The Raman profile (second Raman spectrum 52) shown by the dashed-dotted line in FIG. 9 shows the Raman spectrum measured in the second square region 15. Using the second Raman spectrum 52, the wave number ν2 of the peak corresponding to the folded mode of the longitudinal optical branch is obtained. Similarly, using the second Raman spectrum 52, the full width at half maximum Δ2 of the peak corresponding to the folded mode of the longitudinal optical branch is obtained. Specifically, the full width at half maximum Δ2 is the full width at half maximum (FWHM). Each of the wave number ν2 and the full width at half maximum Δ2 is obtained at 100 measurement positions in the second square region 15. In the second square region 15, the average value of the wave number ν2 is the second wave number. In the second square region 15, the average value of the full width at half maximum Δ2 is the second full width at half maximum.

[0047] In the silicon carbide substrate 10 according to the present embodiment, the absolute value of the difference between the first wave number and the second wave number is 0.2 cm -1 or less, and the absolute value of the difference between the first full width at half maximum and the second full width at half maximum is 0.25 cm -1The following holds. The absolute value of the difference between the first frequency and the second frequency may be 0.18 cm or less, or may be 0.16 cm or less. The lower limit of the absolute value of the difference between the first frequency and the second frequency is not particularly limited, and may be, for example, 0.14 cm or more. -1 or less, or may be 0.16 cm -1 or less. The lower limit of the absolute value of the difference between the first frequency and the second frequency is not particularly limited, and may be, for example, 0.14 cm -1 or more.

[0048] The absolute value of the difference between the first full width at half maximum and the second full width at half maximum may be 0.23 cm or less, or may be 0.21 cm or less. The lower limit of the absolute value of the difference between the first full width at half maximum and the second full width at half maximum is not particularly limited, and may be, for example, 0.20 cm or more. -1 or less, or may be 0.21 cm -1 or less. The lower limit of the absolute value of the difference between the first full width at half maximum and the second full width at half maximum is not particularly limited, and may be, for example, 0.20 cm -1 or more. The frequency of the peak corresponding to the folding mode of the vertical optical branching and the full width at half maximum of the peak change depending on the stress in the measurement region. When the polishing damage is small, the absolute value of the difference between the first frequency and the second frequency and the absolute value of the difference between the first full width at half maximum and the second full width at half maximum become small. In other words, by defining the absolute value of the difference between the first frequency and the second frequency and the absolute value of the difference between the first full width at half maximum and the second full width at half maximum, the degree of polishing damage can be quantified.

[0049] Next, a method for manufacturing the silicon carbide substrate 10 according to the present embodiment will be described. FIG. 10 is a flowchart schematically showing the method for manufacturing the silicon carbide substrate 10 according to the present embodiment. As shown in FIG. 10, the method for manufacturing the silicon carbide substrate 10 according to the present embodiment mainly includes a step of preparing a silicon carbide single crystal substrate 100 (S10: FIG. 10), a step of mechanically polishing the silicon carbide single crystal substrate 100 (S20: FIG. 10), a step of etching the silicon carbide single crystal substrate 100 (S30: FIG. 10), a step of chemically mechanically polishing the silicon carbide single crystal substrate 100 (S40: FIG. 10), and a step of cleaning the silicon carbide single crystal substrate 100 (S50: FIG. 10).

[0050] First, a step of preparing a silicon carbide single crystal substrate 100 (S10: FIG. 10) is performed. Specifically, for example, an ingot composed of a polytype 4H silicon carbide single crystal is formed by the sublimation method. After the ingot is shaped, the ingot is sliced by a wire saw device. Thereby, the silicon carbide single crystal substrate 100 is cut out from the ingot.

[0051] The silicon carbide single crystal substrate 100 is composed of polytype 4H hexagonal silicon carbide. The silicon carbide single crystal substrate 100 has a first main surface 1 and a second main surface 2 on the opposite side of the first main surface 1. The first main surface 1 is a surface that is offset by 4° or less in the <11-20> direction with respect to, for example, the {0001} plane. Specifically, the first main surface 1 is a surface that is offset by an angle of about 4° or less with respect to, for example, the (0001) plane. The second main surface 2 is a surface that is offset by an angle of about 4° or less with respect to, for example, the (000-1) plane.

[0052] As shown in FIG. 11, the silicon carbide substrate 10 has a first main surface 1, a second main surface 2, a plurality of screw dislocations 13, and a silicon carbide region 22. The plurality of screw dislocations 13 are continuous with each of the first main surface 1 and the second main surface 2. From another perspective, the plurality of screw dislocations 13 penetrate the silicon carbide region 22 from the first main surface 1 to the second main surface 2. As described above, a silicon carbide single crystal substrate 100 having a first main surface 1 and a second main surface 2 on the opposite side of the first main surface 1 is prepared.

[0053] Next, a step of mechanically polishing the silicon carbide single crystal substrate 100 (S20: FIG. 10) is performed. Specifically, the first main surface 1 is arranged to face a surface plate (not shown). Next, a slurry is introduced between the first main surface 1 and the surface plate. The slurry contains, for example, diamond abrasive grains. The diameter of the diamond abrasive grains is, for example, 1 μm or more and 3 μm or less. A load is applied to the first main surface 1 by the surface plate. As described above, the silicon carbide single crystal substrate 100 is mechanically polished on the first main surface 1.

[0054] As shown in FIG. 12, in the process of mechanically polishing the silicon carbide single crystal substrate 100, a damage layer 23 is formed on the first main surface 1. In a portion where there is a screw dislocation 13, the damage layer 23 is more likely to be formed compared to a normal crystal portion without the screw dislocation 13. Therefore, the thickness of the damage layer 23 in the portion along the screw dislocation 13 becomes larger than the thickness of the damage layer 23 in the region along the area without the screw dislocation 13. From another perspective, the damage layer 23 is formed so as to erode the silicon carbide region 22 along the extending direction of the screw dislocation 13.

[0055] Next, a process of etching the silicon carbide single crystal substrate 100 (S30: FIG. 10) is performed. As shown in FIG. 13, in the process of etching the silicon carbide single crystal substrate 100, the damage layer 23 formed in the mechanical polishing process is removed. After the damage layer 23 is removed from the first main surface 1, pits 11 are formed on the first main surface 1. The pits 11 are continuous with the screw dislocation 13.

[0056] The silicon carbide single crystal substrate 100 may be etched in the gas phase or in the liquid phase. Preferably, the process of etching the silicon carbide single crystal substrate 100 is performed by immersing the damage layer 23 in an etching solution. The etching solution contains, for example, potassium hydroxide (KOH), potassium permanganate (KMnO4), and pure water. The volume ratio of the etching solution is, for example, KOH:KMnO4:pure water = 5 to 15:1 to 3:30 to 40.

[0057] The process of etching the silicon carbide single crystal substrate 100 is performed, for example, at 400°C or lower. The process of etching the silicon carbide single crystal substrate 100 may be performed, for example, at 350°C or lower, or at 300°C or lower. Specifically, the temperature of the etching solution is, for example, 60°C or higher and 70°C or lower. The etching amount is, for example, about 1 μm or more and 5 μm or less. The process of etching the silicon carbide single crystal substrate 100 is performed after the process of mechanically polishing the silicon carbide single crystal substrate 100.

[0058] Next, a step of chemically and mechanically polishing the silicon carbide single crystal substrate 100 (S40: FIG. 10) is performed. First, the conditions for CMP (Chemical Mechanical Polishing) are determined. The CMP conditions are set such that a balance is achieved between the mechanical and chemical elements. Specifically, while fixing the size of the abrasive grains for CMP and changing the concentration of the oxidizing agent, the polishing rate of the silicon carbide single crystal substrate 100 and the surface roughness (Sa) of the first main surface 1 of the silicon carbide single crystal substrate 100 are measured.

[0059] Specifically, on the first main surface 1, CMP is performed on the silicon carbide single crystal substrate 100 using the abrasive grains and the oxidizing agent. For example, the silicon carbide single crystal substrate 100 is held by a polishing head (not shown) such that the first main surface 1 faces a surface plate (not shown). The abrasive grains are, for example, colloidal silica. The average particle size of the abrasive grains is 20 nm. The working surface pressure is, for example, 400 g / cm 2 . The rotational speed of the surface plate is, for example, 60 rpm. The rotational speed of the polishing head is 60 rpm. The oxidizing agent is, for example, an aqueous solution of aluminum nitrate. The oxidizing agent concentration is, for example, 5%, 10%, 15%, 20%, and 25%. Note that the oxidizing agent concentration is a value obtained by dividing the mass of the solute (aluminum nitrate) by the total mass of the solute (aluminum nitrate) and the solvent (water).

[0060] FIG. 14 is a diagram showing the relationship between each of the polishing rate and the surface roughness and the oxidizing agent concentration. The vertical axis on the left side is the polishing rate. The vertical axis on the right side is the surface roughness of the first main surface 1. The horizontal axis is the oxidizing agent concentration.

[0061] In FIG. 14, the white squares are data on the polishing rate. The solid line is a line obtained by approximating the values of the polishing rate with a quadratic curve (polynomial). A quadratic curve is a curve represented by a quadratic equation. The relationship between the polishing rate and the oxidizing agent concentration is approximated by a downwardly convex quadratic curve. In FIG. 14, the white circles are data on the surface roughness (Sa) of the first main surface 1. The broken line is a line obtained by approximating the values of the surface roughness of the first main surface 1 with a quadratic curve (the first quadratic curve). The relationship between the surface roughness of the first main surface 1 and the oxidizing agent concentration is approximated by an upwardly convex quadratic curve.

[0062] The concentration of the oxidizing agent is determined so as to be within a range where the surface roughness is 1.5 times or less the minimum value of the first quadratic curve. As shown in FIG. 14, the minimum value of the first quadratic curve indicated by the broken line is 0.09 nm. 1.5 times the minimum value is 0.135 nm. Therefore, the concentration of the oxidizing agent is determined within a range where the surface roughness is 0.135 nm or less. Specifically, the concentration of the oxidizing agent is determined, for example, within the range of 8% or more and 16% or less. Preferably, the concentration of the oxidizing agent is determined so as to be within a range where the surface roughness is 1.3 times or less the minimum value of the first quadratic curve.

[0063] Also, the concentration of the oxidizing agent is determined within a range where the polishing rate of the single-crystalline silicon carbide substrate 100 is 0.2 μm / hour or more. As shown in FIG. 14, the concentration of the oxidizing agent at which the polishing rate of the single-crystalline silicon carbide substrate 100 is 0.2 μm / hour or more is, for example, within the range of 5% or more and 22% or less. That is, the concentration of the oxidizing agent that is within a range where the surface roughness is 1.5 times or less the minimum value of the first quadratic curve and at which the polishing rate of the single-crystalline silicon carbide substrate 100 is 0.2 μm / hour or more is, for example, within the range of 8% or more and 16% or less.

[0064] As shown in FIG. 14, the minimum value of the first quadratic curve indicated by the broken line is, for example, 0.15 nm or less. The minimum value of the first quadratic curve indicated by the broken line may be, for example, 0.13 nm or less, or may be 0.11 nm or less.

[0065] FIG. 15 is a diagram showing the relationship between each of the polishing rate and the surface roughness and the abrasive grain size. The vertical axis on the left side is the polishing rate. The vertical axis on the right side is the surface roughness (Sa) of the first main surface 1. The horizontal axis is the abrasive grain size (the diameter of the abrasive grains).

[0066] In FIG. 15, the white squares represent the data of the polishing rate. The solid line is the line obtained by approximating the values of the polishing rate by a power function. When the abrasive grain diameter is less than 6 nm, the polishing rate decreases rapidly. When the abrasive grain diameter is 6 nm or more, the polishing rate does not change much. In FIG. 15, the white circles represent the data of the surface roughness of the first main surface 1. The broken line is the line obtained by approximating the values of the surface roughness of the first main surface 1 by a quadratic curve (second quadratic curve). The relationship between the surface roughness of the first main surface 1 and the abrasive grain diameter is approximated by a downward-convex quadratic curve.

[0067] The diameter of the abrasive grains may be determined so as to be within a range where the surface roughness is 1.5 times or less the minimum value of the second quadratic curve. As shown in FIG. 15, the minimum value of the second quadratic curve indicated by the broken line is 0.09 nm. 1.5 times the minimum value is 0.135 nm. Therefore, the diameter of the abrasive grains is determined within a range where the surface roughness is 0.135 nm or less. Specifically, the diameter of the abrasive grains is determined, for example, within a range of 30 nm or less. Preferably, the diameter of the abrasive grains is determined so as to be within a range where the surface roughness is 1.3 times or less the minimum value of the second quadratic curve.

[0068] As described above, the oxidant concentration and the diameter of the abrasive grains are determined. The oxidant concentration is, for example, 10%. The diameter of the abrasive grains is, for example, 20 nm. Using the above conditions, CMP is performed on the first main surface 1 of the silicon carbide single crystal substrate 100. Note that the CMP of the silicon carbide single crystal substrate 100 is performed after the step of etching the silicon carbide single crystal substrate 100.

[0069] Next, a step of cleaning the silicon carbide single crystal substrate 100 (S50: FIG. 10) is performed. The step of cleaning the silicon carbide single crystal substrate 100 includes, for example, a sulfuric acid overcleaning step, an ammonia overcleaning step, a hydrochloric acid overcleaning step, and a hydrofluoric acid cleaning step.

[0070] First, a sulfuric acid-peroxide washing process is performed. Sulfuric acid-peroxide is a solution in which sulfuric acid, hydrogen peroxide solution, and ultrapure water are mixed. As the sulfuric acid, for example, concentrated sulfuric acid with a mass percentage concentration of 96% can be used. As the hydrogen peroxide solution, for example, hydrogen peroxide solution with a mass percentage concentration of 30% can be used. The same applies to the hydrogen peroxide solution used in the subsequent processes. The volume ratio of sulfuric acid, hydrogen peroxide solution, and ultrapure water contained in sulfuric acid-peroxide is, for example, from 10 (sulfuric acid):1 (hydrogen peroxide solution):1 (ultrapure water) to 10 (sulfuric acid):3 (hydrogen peroxide solution):1 (ultrapure water).

[0071] Next, an ammonia-peroxide washing process is carried out. Ammonia-peroxide is a solution in which an aqueous ammonia solution, hydrogen peroxide solution, and ultrapure water are mixed. As the aqueous ammonia solution, for example, an aqueous ammonia solution with a mass percentage concentration of 28% can be used. The volume ratio of the aqueous ammonia solution, hydrogen peroxide solution, and ultrapure water contained in ammonia-peroxide is, for example, from 1 (aqueous ammonia solution):1 (hydrogen peroxide solution):5 (ultrapure water) to 1 (aqueous ammonia solution):1 (hydrogen peroxide solution):10 (ultrapure water).

[0072] Next, a hydrochloric acid-peroxide washing process is carried out. Hydrochloric acid-peroxide is a solution in which hydrochloric acid, hydrogen peroxide solution, and ultrapure water are mixed. As the hydrochloric acid, for example, concentrated hydrochloric acid with a mass percentage concentration of 98% can be used. The volume ratio of hydrochloric acid, hydrogen peroxide solution, and ultrapure water contained in hydrochloric acid-peroxide is, for example, from 1 (hydrochloric acid):1 (hydrogen peroxide solution):5 (ultrapure water) to 1 (hydrochloric acid):1 (hydrogen peroxide solution):10 (ultrapure water).

[0073] Next, a hydrofluoric acid washing process is carried out. The concentration of hydrofluoric acid in the mixed liquid of hydrofluoric acid and ultrapure water is, for example, 10% or more and 40% or less. The temperature of the hydrofluoric acid is, for example, room temperature. Thus, the silicon carbide substrate 10 according to the present embodiment is manufactured (see FIG. 1).

[0074] FIG. 16 is a schematic cross-sectional view showing the configuration of the silicon carbide substrate 10 according to the present embodiment. As shown in FIG. 16, in the silicon carbide substrate 10 according to the present embodiment, pits 11 having a maximum diameter of 1 μm or more and 10 μm or less are hardly present. Specifically, when observing the screw dislocations 13 and the pits 11 on the first main surface 1, the ratio of the number of pits 11 divided by the number of screw dislocations 13 is 1% or less.

[0075] Next, the operation and effect of the silicon carbide substrate 10 according to the present embodiment will be described. As shown in FIG. 12, when the silicon carbide single crystal substrate 100 is mechanically polished, a damage layer 23 is formed on the first main surface 1. The damage layer 23 is a portion where the crystal structure of silicon carbide is broken and becomes amorphous. In the damage layer 23, the stress is higher compared to the silicon carbide region 22 other than the damage layer 23. After the silicon carbide single crystal substrate 100 is mechanically polished, CMP is performed on the silicon carbide single crystal substrate 100. In CMP, mechanical elements and chemical elements act.

[0076] FIG. 17 is a schematic cross-sectional view showing the configuration of the silicon carbide single crystal substrate 100 after CMP when the chemical element is dominant. The portion of the damage layer 23 having the screw dislocation 13 is easily eroded by the chemical component of CMP. Therefore, pits 11 are likely to be formed in the portion having the screw dislocation 13 (see FIG. 17).

[0077] FIG. 18 is a schematic cross-sectional view showing the configuration of the silicon carbide single crystal substrate 100 after CMP when the mechanical element is dominant. When the mechanical element is dominant, the chemical element becomes relatively weak. Therefore, the portion of the damage layer 23 having the screw dislocation 13 is not easily eroded by the chemical component of CMP. On the other hand, since the mechanical element is relatively strong, the damage layer 23 remains in the portion having the screw dislocation 13. As a result, pits 11 are hardly formed on the first main surface 1 (see FIG. 18). Visually, the first main surface 1 becomes almost flat.

[0078] When forming a silicon carbide layer by epitaxial growth on the first main surface 1 of the silicon carbide substrate 10, hydrogen etching of the silicon carbide substrate 10 is performed on the first main surface 1. The damage layer 23 remaining in the portion with the screw dislocation 13 is easily removed by hydrogen etching.

[0079] FIG. 19 is a schematic cross-sectional view showing the structure of the silicon carbide substrate 10 after hydrogen etching is performed on the silicon carbide substrate 10 after CMP when mechanical elements are dominant. As shown in FIG. 19, the damage layer 23 remaining in the portion with the screw dislocation 13 is removed by hydrogen etching. As a result, a large number of pits 11 are formed on the first main surface 1 of the silicon carbide substrate 10. Thereafter, when a silicon carbide epitaxial layer is formed by epitaxial growth on the first main surface 1, a large number of pits 11 also remain on the surface of the silicon carbide epitaxial layer.

[0080] The silicon carbide substrate 10 according to the present embodiment is formed using a CMP process in which the balance between mechanical elements and chemical elements is achieved. Therefore, in the CMP process, the pits 11 are removed without forming the damage layer 23. As a result, a silicon carbide substrate 10 with the damage layer 23 and the pits 11 suppressed is obtained (see FIG. 16).

[0081] FIG. 20 is a schematic cross-sectional view showing the structure of the silicon carbide substrate 10 after hydrogen etching is performed on the silicon carbide substrate 10 after CMP in which the balance between mechanical elements and chemical elements is achieved. As shown in FIG. 20, even after hydrogen etching, since the damage layer 23 does not remain, almost no pits 11 are formed on the first main surface 1. That is, even when hydrogen etching of the silicon carbide substrate 10 is performed on the first main surface 1, it is possible to suppress the formation of pits 11 on the first main surface 1 of the silicon carbide substrate 10. Therefore, when a silicon carbide epitaxial layer is formed by epitaxial growth on the first main surface 1, it is possible to suppress the formation of pits 11 on the surface of the silicon carbide epitaxial layer.

Example

[0082] (Sample Preparation) First, silicon carbide substrates 10 according to Samples 1 to 3 were prepared. The silicon carbide substrates 10 according to Samples 1 and 2 were used as comparative examples. The silicon carbide substrate 10 according to Sample 3 was used as an example. For the silicon carbide substrate 10 according to Sample 3, a step of etching the single-crystalline silicon carbide substrate 100 (S30: Fig. 10) was performed. On the other hand, for the silicon carbide substrates 10 according to Samples 1 and 2, the step of etching the single-crystalline silicon carbide substrate 100 (S30: Fig. 10) was not performed.

[0083] Regarding the silicon carbide substrate 10 according to Sample 1, the dominant factor in the step of chemically mechanical polishing the single-crystalline silicon carbide substrate 100 (S40: Fig. 10) was the mechanical factor. Regarding the silicon carbide substrate 10 according to Sample 2, the dominant factor in the step of chemically mechanical polishing the single-crystalline silicon carbide substrate 100 (S40: Fig. 10) was the chemical factor. Regarding the silicon carbide substrate 10 according to Sample 3, the dominant factors in the step of chemically mechanical polishing the single-crystalline silicon carbide substrate 100 (S40: Fig. 10) were made to be the mechanical factor and the chemical factor to the same extent. In each of the silicon carbide substrates 10 according to Samples 1 to 3, the screw dislocation density was 390 pieces / cm 2 , 420 pieces / cm 2 and 350 pieces / cm 2 respectively. (Evaluation Method) Using the X-ray topography method, the density of screw dislocations 13 on the first main surface 1 of the silicon carbide substrates 10 according to Samples 1 to 3 was measured. Using a defect inspection device, the density of pits 11 on the first main surface 1 of the silicon carbide substrates 10 according to Samples 1 to 3 was measured. The maximum diameter (diameter) of the pits 11 is 1 μm or more and 10 μm or less.

[0084] Using a white interference microscope, the surface roughness of the first main surface 1 of the silicon carbide substrate 10 according to Samples 1 to 3 was measured. The surface roughness of the first main surface 1 was defined as the arithmetic mean roughness (Sa). The measurement range of the arithmetic mean roughness (Sa) was a square region of 255 μm × 255 μm. The center of the diagonal of the square region was made the center of the first main surface 1. One side of the square region was made parallel to the extending direction of the first flat.

[0085] Using Raman spectroscopy, the Raman spectrum of the silicon carbide substrate 10 was measured in each of the first square region 14 and the second square region 15 of the first main surface 1 of the silicon carbide substrate 10 according to Samples 1 to 3. The first square region 14 is a region including the screw dislocation 13. The first square region 14 is a square region of 200 μm × 200 μm. The number of measurement points is 100 points. The second square region 15 is a region not including the screw dislocation 13. The second square region 15 is a square region of 200 μm × 200 μm. The number of measurement points is 100 points. Using the Raman spectrum, the average value of Δν(Ne) and the average value of the full width at half maximum (FWHM) of the peak were obtained.

[0086] Δν(Ne) is a value obtained by subtracting the wave number of the peak of the Raman spectrum of neon from the wave number of the peak corresponding to the folded mode of the longitudinal optical branch of polytype 4H silicon carbide. Based on the wave number indicating the peak of the Raman spectrum of neon, the wave number of the peak corresponding to the folded mode of the longitudinal optical branch of silicon carbide was obtained. The full width at half maximum (FWHM) of the peak is the full width at half maximum of the peak corresponding to the folded mode of the longitudinal optical branch of polytype 4H silicon carbide.

[0087] Next, a silicon carbide epitaxial layer was formed on the first main surface 1 by epitaxial growth. Using a defect measurement device, the density of the pits 11 on the surface of the silicon carbide epitaxial layer was measured. The maximum diameter of the pits 11 is 1 μm or more and 10 μm or less. (Evaluation Results)

[0088] [Table 1]

[0089] As shown in Table 1, the density of the pits 11 on the first main surface 1 of the silicon carbide substrates 10 according to Samples 1 to 3 was 12 pieces / cm 2 , 0.7 pieces / cm 2 and 1.6 pieces / cm 2 respectively. The values obtained by dividing the density of the pits 11 on the first main surface 1 of the silicon carbide substrates 10 according to Samples 1 to 3 by the density of the screw dislocations 13 were 3.0%, 0.2% and 0.4% respectively. The surface roughness (Sa) on the first main surface 1 of the silicon carbide substrates 10 according to Samples 1 to 3 was 0.26 nm, 0.19 nm and 0.09 nm respectively.

[0090] Δν(Ne) in the first square region 14 of the first main surface 1 of the silicon carbide substrates 10 according to Samples 1 to 3 was -44.05 cm -1 , -44.25 cm -1 and -44.33 cm -1 respectively. Δν(Ne) in the second square region 15 of the first main surface 1 of the silicon carbide substrates 10 according to Samples 1 to 3 was -44.21 cm -1 , -44.48 cm -1 and -44.49 cm -1 respectively. The differences between Δν(Ne) in the first square region 14 and Δν(Ne) in the second square region 15 of the first main surface 1 of the silicon carbide substrates 10 according to Samples 1 to 3 were 0.16 cm -1 , 0.23 cm -1 and 0.16 cm -1 respectively.

[0091] The full width at half maximum of the peak in the first square region 14 of the first main surface 1 of the silicon carbide substrates 10 according to Samples 1 to 3 was 2.62 cm -1 , 2.74 cm -1 and 2.58 cm -1 respectively. Δν(Ne) in the second square region 15 of the first main surface 1 of the silicon carbide substrates 10 according to Samples 1 to 3 was 2.33 cm -1 , 2.28 cm -1 and 2.35 cm -1It was. The differences between the full width at half maximum in the first square region 14 and the full width at half maximum in the second square region 15 of the silicon carbide substrate 10 according to Samples 1 to 3 were 0.29 cm -1 , 0.46 cm -1 and 0.23 cm -1 respectively.

[0092] As shown in Table 1, the density of pits 11 on the surface of the silicon carbide epitaxial layer formed by epitaxial growth on the first main surface 1 of the silicon carbide substrate 10 according to Samples 1 to 3 was 375 pits / cm 2 , 364 pits / cm 2 and 2.5 pits / cm 2 respectively. From the above results, it was confirmed that the silicon carbide substrate 10 according to Sample 3 can suppress the formation of pits 11 after epitaxial growth as compared with the silicon carbide substrates 10 according to Samples 1 and 2 respectively.

[0093] The embodiments and examples disclosed this time should be considered illustrative in all respects and not restrictive. The scope of the present invention is shown by the claims rather than the above description, and it is intended that all meanings equivalent to the claims and all modifications within the scope are included.

Explanation of Signs

[0094] 1 First main surface, 2 Second main surface, 3 First flat, 4 Arc-shaped part, 5 Outer peripheral surface, 6 First screw dislocation, 7 Second screw dislocation, 10 Silicon carbide substrate, 11 Pit, 13 Screw dislocation, 14 First square region, 15 Second square region, 22 Silicon carbide region, 23 Damage layer, 30 Raman spectrometer, 31 Objective lens, 32 Light source, 33 Spectrometer, 34 Stage, 35 Beam splitter, 36 Incident light, 38 Detector, 41 First peak, 42 Second peak, 43 Third peak, 44 Fourth peak, 51 First Raman spectrum, 52 Second Raman spectrum, 61, 62, 63 Arrows, 100 Silicon carbide single crystal substrate, 101 First direction, 102 Second direction, A Maximum diameter, D First depth, W First diameter.

Claims

1. A silicon carbide substrate comprising a first main surface and a second main surface on the opposite side of the first main surface, The silicon carbide substrate includes screw dislocations and pits having a maximum diameter in a direction parallel to the first main surface of 1 μm or more and 10 μm or less, When observing the screw dislocations and the pits on the first main surface, the ratio of the number of the pits divided by the number of the screw dislocations is 1% or less, The surface roughness of the first main surface is 0.15 nm or less, The first main surface has a first square region and a second square region, In the first square region including the screw dislocations and having a side length of 200 μm, the average value of the wave numbers showing peaks corresponding to the folded modes of the longitudinal optical branches of the Raman spectrum of silicon carbide is defined as the first wave number, In the second square region not including the screw dislocations and having a side length of 200 μm, the average value of the wave numbers showing peaks corresponding to the folded modes of the longitudinal optical branches of the Raman spectrum of silicon carbide is defined as the second wave number, In the first square region, the average value of the half-widths at half maximum of the peaks corresponding to the folded modes of the longitudinal optical branches of the Raman spectrum of silicon carbide is defined as the first half-width, and In the second square region, when the average value of the half-widths at half maximum of the peaks corresponding to the folded modes of the longitudinal optical branches of the Raman spectrum of silicon carbide is defined as the second half-width, The absolute value of the difference between the first wave number and the second wave number is 0.2 cm -1 or less, and the absolute value of the difference between the first half-width and the second half-width is 0.25 cm -1 or less, The screw dislocations include a first screw dislocation connected to the pits and a second screw dislocation not connected to the pits. A silicon carbide substrate.

2. The silicon carbide substrate according to claim 1, wherein the ratio of the number of the pits divided by the number of the screw dislocations is 0.5% or less.

3. The ratio obtained by dividing the number of the pits by the number of the screw dislocations is 0.4% or less. The silicon carbide substrate according to claim 1.

4. The surface roughness of the first main surface is 0.1 nm or less. The silicon carbide substrate according to any one of claims 1 to 3.

5. The diameter of the first main surface is 150 mm or more. The silicon carbide substrate according to any one of claims 1 to 4.

6. The surface density of the screw dislocations on the first main surface is 100 cm -2 or more and 5000 cm -2 or less. The silicon carbide substrate according to any one of claims 1 to 5.

7. A step of preparing a silicon carbide substrate according to any one of claims 1 to 6, and a step of forming an epitaxial growth on the silicon carbide substrate. A method for manufacturing a silicon carbide epitaxial substrate.

8. A step of preparing a silicon carbide single crystal substrate having a first main surface and a second main surface on the opposite side of the first main surface, a step of mechanically polishing the silicon carbide single crystal substrate on the first main surface, a step of etching the silicon carbide single crystal substrate after the step of mechanically polishing the silicon carbide single crystal substrate, and a step of chemically mechanically polishing the silicon carbide single crystal substrate using abrasive grains and an oxidizing agent on the first main surface after the step of etching the silicon carbide single crystal substrate. In the step of mechanically polishing the silicon carbide single crystal substrate, a damaged layer is formed on the first main surface, and in the step of etching the silicon carbide single crystal substrate, the damaged layer is removed. In the step of chemically and mechanically polishing the silicon carbide single crystal substrate, when the relationship between the surface roughness of the first main surface and the concentration of the oxidizing agent is approximated by a first quadratic curve with the surface roughness of the first main surface as the vertical axis and the concentration of the oxidizing agent as the horizontal axis, the concentration of the oxidizing agent is within the range of the surface roughness that is 1.5 times or less of the minimum value of the first quadratic curve, and the polishing rate of the silicon carbide single crystal substrate is 0.2 μm / hour or more. A method for manufacturing a silicon carbide substrate, wherein in the step of chemically and mechanically polishing the silicon carbide single crystal substrate, when the relationship between the surface roughness of the first main surface and the diameter of the abrasive grains is approximated by a second quadratic curve with the surface roughness of the first main surface as the vertical axis and the diameter of the abrasive grains as the horizontal axis, the diameter of the abrasive grains is within the range of the surface roughness that is 1.5 times or less of the minimum value of the second quadratic curve.

9. The method for manufacturing a silicon carbide substrate according to claim 8, wherein the step of etching the silicon carbide single crystal substrate is performed at 400°C or lower.

10. The method for manufacturing a silicon carbide substrate according to claim 8 or claim 9, wherein the minimum value of the first quadratic curve is 0.15 nm or less.

11. The method for manufacturing a silicon carbide substrate according to any one of claims 8 to 10, wherein the abrasive grains are colloidal silica.

12. The method for manufacturing a silicon carbide substrate according to any one of claims 8 to 11, wherein the step of etching the silicon carbide single crystal substrate is performed by immersing the damaged layer in a solution.

13. The method for manufacturing a silicon carbide substrate according to claim 12, wherein the solution contains potassium permanganate and potassium hydroxide.

Images