Silicon carbide wafer and grinding method thereof

Abstract

The present invention describes a method for grinding a silicon carbide wafer to improve its geometric curvature. First, a single crystal is cut into several wafers, each of which has a silicon face, i.e., a first surface. The opposite side is a carbon face, called a second surface. Then, the silicon face of the wafer is facing down and placed on a grinding carrier to perform the first grinding. It should be emphasized that there is a supporting carrier between the wafer and the grinding carrier. The carrier can be a concave or convex frame. After grinding the carbon face, the wafer is removed and the carrier is removed, and the wafer will appear convex or concave on the carbon face. Later, the wafer is turned upside down, with the carbon face facing down, and placed on a flat carrier. Finally, the silicon face is polished, which is the second grinding. In this way, using the supporting carrier of the present invention, the geometric curvature of the wafer can be improved.

Description

Technical Field

[0001] The present invention relates to a silicon carbide wafer, and in particular to a silicon carbide wafer and a grinding method thereof, which utilizes a preset external bending deformation of the wafer to buffer or offset the influence of the internal residual stress of the wafer on the curvature of the wafer, and finally improves its geometric curvature. Background Art

[0002] Silicon carbide ingots are generally formed by physical vapor transport, where solid silicon carbide raw materials are volatilized by induction heating in a high-temperature furnace at about 2450 degrees Celsius, and then deposited on the surface of the seed crystal, growing and thickening. Figure 1 As shown, in the conventional crystal growth process, a seed crystal 106 is placed at the ceiling end of a high temperature furnace 102, and a silicon carbide powder 110 is placed at the bottom end of the high temperature furnace 102 as a solid evaporation source, and the high temperature furnace 102 is heated by an induction coil 104. During the process, the seed crystal 106 receives the raw material (silicon carbide powder 110) that is solidified again from the gaseous state, and slowly forms a semiconductor material on the surface of the seed crystal 106 until a crystal ingot 108 of a desired size is obtained. The crystal ingot 108 may have different crystal structures depending on the manufacturing method, the manufacturing raw material, and the crystal orientation of the seed crystal. For example, the crystal ingot of silicon carbide includes 3C-silicon carbide, 4H-silicon carbide, 6H-silicon carbide, etc. 3C-silicon carbide belongs to the cubic system, while 4H-silicon carbide and 6H-silicon carbide belong to the hexagonal system.

[0003] In the crystal growth process, the ingot 108 grows in a high temperature environment of 2300 degrees Celsius to 2500 degrees Celsius. Among them, the end of the growing ingot 108 close to the seed crystal 106 is called the seed end 108B (seed end), and the end of the ingot 108 away from the seed crystal 106 is called the dome end 108A (dome end). The seed end 108B and the dome end 108A may have a temperature difference of tens to hundreds of degrees Celsius due to different temperature distributions in the furnace. In this case, the interior of the ingot 108 may show residual stress when the crystal growth is completed and the temperature drops to room temperature. Taking silicon carbide as an example, the seed end 108B is the silicon surface, and the dome end 108A is the carbon surface. The silicon surface may develop residual compressive stress, and the carbon surface may show residual tensile stress.

[0004] More specifically, after the growth of the ingot 108 is completed, the ingot 108 is cooled to room temperature by furnace cooling or other methods. When the ingot 108 is cooled to below the plastic-elastic transition temperature zone, the shrinkage deformation caused by the continued cooling of the ingot 108 gradually cannot be released by real-time dislocation slip, generation, combination, strain relaxation in the form of plastic deformation, or instantaneous thermal stress re-distribution. Gradually, the accumulated shrinkage deformation differences at different time sequences and different positions cause residual stress. If the ingot 108 is considered as a simple elastic material property, the thermal shrinkage of the ingot roughly conforms to the following formula:

[0005] ε=kΔΤ

[0006] In the above formula, ε is strain, k is the coefficient of thermal expansion, and ΔT is the temperature difference. When the growth of the ingot 108 is completed and the cooling has passed the plastic-elastic transition temperature zone, when the temperature is lowered again, the ingot 108 can be simply regarded as an elastic material, and the seed end 108B and the dome end 108A start to cool down from different temperatures. For example, the seed end 108B may be cooled from 1800 degrees Celsius to 20 degrees, and the dome end 108A may be cooled from 1900 degrees Celsius to 20 degrees. In this case, the temperature difference of the dome end 108A is larger, and it shrinks more. Therefore, its contraction must be restrained and hindered by the seed end 108B, and tensile stress is generated at the dome end. In contrast, the seed end 108B exhibits compressive stress.

[0007] Figure 2A and Figure 2B According to some embodiments, a crystal ingot obtained through a crystal growth process and a cross-sectional schematic diagram of a cutting process of the crystal ingot are shown. Figure 2A As shown in FIG. 1 , the seed end 108B of the crystal ingot 108 obtained through the crystal growth process will exhibit residual compressive stress, and the round end 108A of the crystal ingot 108 will exhibit residual tensile stress. Figure 2B As shown, the obtained ingot 108 is sliced ​​to obtain a plurality of wafers. For example, the edges and corners of the ingot 108 are cut into equal-diameter cylinders along the first direction D1, and are ground into rounded corners to prevent the edges and corners of the wafers from breaking due to collision. Next, the ingot 108 is sliced ​​along the second direction D2 to separate the plurality of wafers. The slicing method of the ingot 108 includes cutting with a tool or steel wire in combination with abrasive particles (abrasive particles such as diamond particles). In some cases, the interior of the cut wafer has residual compressive stress and tensile stress like the ingot 108.

[0008] Figure 3A FIG. 1 is a schematic cross-sectional view of the wafer 108W obtained after being cut through the above process. Figure 3AAs shown, in the chip 108W formed by the conventional process, the residual stress causes the basal plane BP (BasalPlane; also called C-plane) formed by the arrangement of atoms in the crystal to bend. The degree of bending of this basal plane BP can be determined by X-ray diffraction according to the angle shift dθ of a specific diffraction angle 2θ. For chips with the same thickness h, the smaller the bending radius γ, the greater the residual stress Stress. Stress = {E.(h / 2)} / {γ.(1-υ)}. E represents the Young's modulus of the chip, and υ is the Poussin's modulus.

[0009] In addition, if Figure 3B As shown, after the wafer 108W is subjected to the grinding and polishing process, the thickness h of the wafer 108W is reduced (for example, reduced by hundreds of micrometers), and the rigidity (rigidity: D = Eh 3 / {12.(1-υ 2 The residual stress inside the chip 108W will be partially released due to the reduced rigidity of the chip, and stress re-distribution will occur, thereby making the base surface BP tend to be flat, which will lead to the appearance of the chip 108W with geometric warping such as bow and / or warp.

[0010] Therefore, how to ensure that all silicon carbide wafers still have an ideal surface flatness after the ingots are cut and polished is an important issue in the physical vapor transport semiconductor material process. Summary of the invention

[0011] The invention provides a grinding method for a silicon carbide wafer, which utilizes a preset external bending deformation of the wafer to buffer or offset the influence of the internal residual stress of the wafer on the wafer curvature, thereby achieving a silicon carbide wafer with a flat appearance.

[0012] An embodiment of the present invention provides a method for grinding a silicon carbide wafer, comprising the following steps. A crystal ingot obtained through a crystal growth process is cut to obtain a silicon carbide wafer, wherein the silicon carbide wafer includes a first surface and a second surface opposite to the first surface. The silicon carbide wafer is placed on a grinding platform for a grinding process, wherein a carrier having a concave or convex surface is included between the grinding platform and the silicon carbide wafer. The grinding process comprises the following steps. The silicon carbide wafer is placed on the concave or convex surface of the carrier so that the first surface contacts the concave or convex surface, and is ground to remove a portion of the second surface of the silicon carbide wafer. The carrier is removed, and the silicon carbide wafer is reversed and placed on a perfectly flat grinding platform without any carrier, and is ground again to remove a portion of the first surface of the silicon carbide wafer.

[0013] In some embodiments, the concave surface of the carrier has a concave depth of 20 μm to 60 μm, or the convex surface of the carrier has a convex height of 20 μm to 60 μm.

[0014] In some embodiments, when the first surface has residual compressive stress and the second surface has residual tensile stress, a bearing with a convex surface should be used. When the first surface has residual tensile stress and the second surface has residual compressive stress, a bearing with a concave surface should be used.

[0015] In some embodiments, after removing the carrier, the silicon carbide wafer is inverted and placed on a flat surface on a grinding stage, so that the second surface contacts the flat surface, and is ground to remove a portion of the first surface of the silicon carbide wafer.

[0016] In some embodiments, the wafer is ground once on both the front and back sides, and when the carrier has a concave surface, the first surface will develop an appearance with a negative curvature, and when the carrier has a convex surface, the first surface will develop an appearance with a positive curvature.

[0017] In some embodiments, the carrier is an adhesive tape including a convex surface, and the adhesive tape includes a first adhesive layer, a second adhesive layer, and a third adhesive layer. The first adhesive layer has a first diameter DA1. The second adhesive layer is located on the first adhesive layer and has a second diameter DA2. The third adhesive layer is located on the second adhesive layer and has a third diameter DA3, wherein DA1>DA2>DA3.

[0018] In some embodiments, the carrier is an adhesive tape including a concave surface, and the adhesive tape includes a first adhesive layer, a second adhesive layer, and a third adhesive layer. The second adhesive layer is located on the first adhesive layer, wherein the second adhesive layer has a first opening to expose a portion of the first adhesive layer. The third adhesive layer is located on the second adhesive layer, wherein the third adhesive layer has a second opening to expose a portion of the second adhesive layer and the first adhesive layer, and the second opening overlaps the first opening.

[0019] In some embodiments, the carrier is formed by screen printing using ink. Screen printing includes the following steps: providing a dot structure on a printed object, wherein the dot structure shields a portion of the printed object and exposes another portion of the printed object. Screen printing is performed so that the ink is printed on another portion of the printed object through the dot structure. Vibrating is performed so that the ink creeps slightly to be flattened, and the surface ink is partially dissolved with a solvent to make it more flattened, and finally a drying step is performed to form the carrier.

[0020] In some embodiments, the mesh structure includes a plurality of shielding portions, and as the plurality of shielding portions are disposed closer to the center of the mesh structure, their widths will decrease while their spacings will increase, and the carrier formed thereby is regarded as a convex surface.

[0021] In some embodiments, the mesh structure includes a plurality of shielding portions, and as the plurality of shielding portions are disposed closer to the center of the mesh structure, their widths will increase while their spacings will decrease, and the carrier formed thereby will be regarded as a concave surface.

[0022] An embodiment of the present invention further provides a silicon carbide wafer manufactured by the above method, wherein the curvature of the silicon carbide wafer is within the range of ±25 μm.

[0023] Based on the above, by setting a carrier with a concave or convex surface on the grinding platform for grinding, the preset external bending deformation of the wafer is used to buffer or offset the influence of the internal residual stress of the wafer on the wafer curvature, thereby achieving a silicon carbide wafer with a flat appearance. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a schematic diagram of a conventional crystal growth apparatus;

[0025] Figure 2A and Figure 2B is a schematic cross-sectional view of a crystal ingot obtained through a crystal growth process and a cutting process of the crystal ingot according to some embodiments;

[0026] Figure 3A and Figure 3B Schematic diagram of the cross-sectional changes of a wafer before and after grinding after cutting according to some embodiments. Figure 3A ), and after thinning ( Figure 3B ), showing that the wafer appearance tends to be warped, but the BP surface (Basal Plane) inside the wafer tends to be flat;

[0027] Figure 4A is a comparison chart of the curvature of the wafer obtained by cutting the wafer in some comparative examples without using the present invention, and the continuous change of the curvature of the wafer by the subsequent grinding and polishing process;

[0028] Figure 4B It is a comparison diagram of the curvature of the wafer obtained by cutting the wafer ingot according to some comparative examples of the present invention, and the continuous change of the curvature of the wafer by the subsequent grinding and polishing process;

[0029] FIG. 5A to FIG. 5H A method for grinding a silicon carbide wafer according to some embodiments of the present invention;

[0030] 6A to 6H A method for grinding a silicon carbide wafer according to other embodiments of the present invention;

[0031] Fig. 7A and Figure 7B are cross-sectional views and exploded top views of carriers according to some embodiments of the present invention;

[0032] Fig. 8A and Figure 8B are cross-sectional views and exploded top views of carriers according to other embodiments of the present invention;

[0033] Fig. 9 A schematic cross-sectional view of screen printing in some embodiments of the present invention;

[0034] FIG. 10A to FIG. 10D FIG. 4 is a top view of a mesh structure according to some embodiments of the present invention.

[0035] Description of Reference Numerals

[0036] 102:High temperature furnace

[0037] 104: Induction coil

[0038] 106: Seed

[0039] 108: Crystal

[0040] 108A: Second surface (dome top)

[0041] 108B: first surface (seed end)

[0042] 108W:Chip

[0043] 110: Silicon carbide powder

[0044] 202: Grinding stage

[0045] 204: Bearing

[0046] 204A: First adhesive layer

[0047] 204B: Second adhesive layer

[0048] 204C: The third adhesive layer

[0049] 204D: Concave depth

[0050] 204H: protrusion height

[0051] 205: Second bearing member

[0052] 300: Printed material

[0053] 302: Network structure

[0054] 302P: Shielded part

[0055] BP: Base plane

[0056] D1: First direction

[0057] D2: Second direction

[0058] DA1: First diameter

[0059] DA2: Second diameter

[0060] DA3: Third diameter

[0061] OP1: First opening

[0062] OP2: Second opening

[0063] SX1, SX2, SX3: Spacing

[0064] W1, W2, W3: Width DETAILED DESCRIPTION

[0065] Reference will now be made in detail to exemplary embodiments of the present invention, examples of which are illustrated in the accompanying drawings. Whenever possible, the same reference numerals are used in the drawings and the description to refer to the same or like parts.

[0066] Figure 4A The following is a comparison chart of the continuous change in curvature of 14 wafers obtained by cutting a certain ingot according to some comparative examples of the present invention, among which there are 7 wafers arranged in an even sequence from the dome end to the seed end, after using the traditional grinding and polishing process. Figure 4A As shown in the chart, the wafer 108W just cut is still thick and rich in residual stress. Through the wire cutting technology, the curvature of the wafer 108W near the round end 108A or the seed end 108B is roughly the same (15.7μm to 18.5μm). After the traditional grinding and polishing process, due to the reduction of the wafer rigidity and the bow / or bending caused by the partial release of stress, the wafer 108W near the seed end 108B will have a larger curvature (-30.8μm after grinding; -47.2μm after polishing), while the wafer 108W near the round end 108A will have a smaller curvature (-11.0μm after grinding; -20.0μm after polishing).

[0067] Figure 4B This is a comparison chart of the continuous change in curvature of 14 chips obtained by cutting the front wafer in accordance with some embodiments of the present invention, after adopting the grinding and polishing process of the present invention, there are a total of 7 chips arranged in an odd sequence from the round end to the seed end. Figure 4BIt is shown that after the wafer 108W is ground and polished by the wafer grinding method of the present invention, the curvature of the wafer 108W at the round end 108A or the wafer 108W near the seed end 108B can be controlled within the range of ±25μm. This is because the preset external bending deformation of the wafer is used in advance to buffer or offset the influence of the residual stress inside the wafer on the curvature of the wafer, thereby achieving a flat appearance. Next, the wafer grinding method of some embodiments of the present invention will be described in detail.

[0068] FIG. 5A to FIG. 5H The present invention provides a method for grinding a silicon carbide wafer according to other embodiments of the present invention.

[0069] refer to Figure 5A , providing a silicon carbide wafer 108W. The silicon carbide wafer 108W is made by Figure 1 The silicon carbide wafer 108W is obtained by cutting the ingot 108 through a crystal growth process or other existing processes. In an embodiment of the present invention, the second surface 108A is a carbon surface with residual tensile stress, and the first surface 108B is a silicon surface with residual compressive stress, but the present invention is not limited thereto. In some other embodiments, depending on different process conditions, the second surface 108A may also be a silicon surface with residual tensile stress, and the first surface 108B may also be a carbon surface with residual compressive stress.

[0070] Next, refer to Figure 5B , is to place the silicon carbide wafer 108W on the grinding stage 202 for a grinding process. In some embodiments, a carrier 204 having a convex surface is included between the grinding stage 202 and the silicon carbide wafer 108W. The convex surface of the carrier 204 has a protrusion height 204H of 20 μm to 60 μm. In some embodiments, the protrusion height 204H of the carrier 204 can be adjusted according to the curvature obtained by performing a general grinding and polishing process on the wafer. The convex surface setting of the carrier 204 can be expected to offset the curvature of wafers of different sorts. For example, referring to Figure 4A According to the graph of FIG. 1 , the wafer near the seed end has a curvature of about -47 μm after a general grinding and polishing process (after partial stress release). Therefore, in some embodiments, the protrusion height 204H of the carrier 204 can be set to about 47 μm to offset the curvature of such wafers.

[0071] like Figure 5BAs shown, during the grinding process, the silicon carbide wafer 108W is attached and placed on the convex surface of the carrier 204, so that the first surface 108B contacts the convex surface, and grinding is performed to remove a portion of the second surface 108A of the silicon carbide wafer 108W. In some embodiments, when the carrier 204 has a convex surface, the first surface 108B will present a negative curvature in order to offset this negative curvature. In some embodiments, since the silicon carbide wafer 108W is placed on the convex surface, there will be a convex portion on the second surface 108A of the silicon carbide wafer 108W. The grinding process can remove the convex portion on the second surface 108A to obtain a Figure 5C The structure shown.

[0072] Next, refer to Figure 5D , is to remove the carrier 204 and remove the silicon carbide wafer 108W. Figure 5D As shown in FIG. 1 , due to the release of some residual stress, the negative curvature of the first surface 108B is slightly offset. Figure 5E Turn the silicon carbide wafer 108W upside down and then Fig. 5F As shown, the silicon carbide wafer 108W is attached and placed on the grinding stage 202. In some embodiments, after removing the carrier 204, the silicon carbide wafer 108W is reversed and placed on the flat surface of the second carrier 205 on the grinding stage 202, so that the second surface 108A of the silicon carbide wafer 108W contacts the flat surface, and grinding is performed to remove a portion of the first surface 108B of the silicon carbide wafer 108W. In some embodiments, when the silicon carbide wafer 108W is reversed and placed on the grinding stage 202, there will be protruding or uneven portions on the first surface 108B of the silicon carbide wafer 108W (such as at the edge). The grinding process can remove the protruding or uneven portions on the first surface 108B to obtain the following. Figure 5G The structure shown.

[0073] Finally, reference Figure 5H After the silicon carbide wafer 108W undergoing the grinding process is removed from the grinding stage 202, the residual stress inside the silicon carbide wafer 108W is partially released due to the reduction in thickness of the silicon carbide wafer 108W, thereby offsetting the negative curvature of the first surface 108B to form a silicon carbide wafer 108W with high flatness. The silicon carbide wafer formed by the above method can effectively control its curvature within the range of ±25μm.

[0074] 6A to 6H A method for grinding a silicon carbide wafer according to some embodiments of the present invention is disclosed.

[0075] refer to Fig. 6A, providing a silicon carbide wafer 108W. The silicon carbide wafer 108W is made by Figure 1 A silicon carbide wafer 108W is obtained by cutting a crystal ingot 108 obtained by a crystal growth process or other existing processes. In some embodiments, the silicon carbide wafer 108W has a first surface 108B and a second surface 108A opposite to the first surface 108B. In the illustrated embodiment, the first surface 108B is a surface having residual compressive stress, and the second surface 108A is a surface having residual tensile stress. In an embodiment of the present invention, the first surface 108B is a silicon surface and the second surface 108A is a carbon surface, but the present invention is not limited thereto. In some other embodiments, depending on the process conditions, the first surface 108B may also be a carbon surface having residual compressive stress, and the second surface 108A may also be a silicon surface having residual tensile stress.

[0076] Next, refer to Figure 6B , is to attach and place the silicon carbide wafer 108W on the grinding stage 202 for a grinding process. In some embodiments, a carrier 204 having a concave surface is included between the grinding stage 202 and the silicon carbide wafer 108W. The concave surface of the carrier 204 has a concave depth 204D of 20 μm to 60 μm. In some embodiments, the concave depth 204D of the carrier 204 can be adjusted according to the curvature obtained by performing a general grinding and polishing process on the wafer. For example, in some embodiments, the wafer near the seed end has a positive curvature after a general grinding and polishing process (after stress release). Therefore, in some embodiments, the concave depth 204D of the carrier 204 can be set to a depth that is expected to offset the positive curvature of such a wafer.

[0077] like Figure 6B As shown, during the grinding process, the silicon carbide wafer 108W is placed on the concave surface of the carrier 204, so that the first surface 108B contacts the concave surface, and grinding is performed to remove a portion of the second surface 108A of the silicon carbide wafer 108W. In some embodiments, when the carrier 204 has a concave surface, the first surface 108B has a positive curvature in order to offset this positive curvature. In some embodiments, since the silicon carbide wafer 108W is placed on the concave surface, there will be a protruding or uneven portion on the second surface 108A of the silicon carbide wafer 108W (such as at the edge). The grinding process can remove the protruding or uneven portion on the second surface 108A to obtain a Figure 6C The structure shown.

[0078] Next, refer to Fig.6D , is to remove the carrier 204 and remove the silicon carbide wafer 108W. Fig.6DAs shown in FIG. 1 , due to the release of some residual stress, the positive curvature of the first surface 108B is slightly offset. Fig. 6E Turn the silicon carbide wafer 108W upside down and then Fig. 6F As shown, the silicon carbide wafer 108W is attached and placed on the grinding stage 202. In some embodiments, after removing the carrier 204, the silicon carbide wafer 108W is reversed and placed on the flat surface of the second carrier 205 on the grinding stage 202, so that the second surface 108A of the silicon carbide wafer 108W contacts the flat surface, and is ground to remove a portion of the first surface 108B of the silicon carbide wafer 108W. In some embodiments, when the silicon carbide wafer 108W is reversed, the first surface 108B of the silicon carbide wafer 108W may have a protruding or uneven portion. The grinding process can remove the protruding or uneven portion on the first surface 108B to obtain a Figure 6G The structure shown.

[0079] Finally, reference Figure 6H After the silicon carbide wafer 108W undergoing the grinding process is removed from the grinding stage 202, the residual stress inside the silicon carbide wafer 108W is partially released due to the reduction in thickness of the silicon carbide wafer 108W, thereby offsetting the positive curvature of the first surface 108B to form a silicon carbide wafer 108W with high flatness. The curvature of the silicon carbide wafer formed by the above method can be effectively controlled within the range of ±25μm.

[0080] Next, the above FIG. 5A to FIG. 5H as well as 6A to 6H The carrier 204 with a concave or convex surface used in the description is described.

[0081] Fig. 7A FIG. 2 is a cross-sectional view of a carrier 204 having a convex surface according to some embodiments of the present invention. Figure 7B For reference Fig. 7A An exploded top view of the carrier 204. Fig. 7A and Figure 7B In some embodiments, the carrier 204 is an adhesive tape including a convex surface, and the adhesive tape includes a first adhesive layer 204A, a second adhesive layer 204B, and a third adhesive layer 204C. The first adhesive layer 204A has a first diameter DA1. The second adhesive layer 204B is located on the first adhesive layer 204A and has a second diameter DA2. The third adhesive layer 204C is located on the second adhesive layer 204B and has a third diameter DA3, wherein DA1>DA2>DA3.

[0082] In the embodiment of the present invention, although the carrier 204 with a convex surface is formed by three layers of adhesive tape, the present invention is not limited thereto. In other embodiments, the carrier 204 may also include one or more than three layers of adhesive tape to form the carrier 204 with a convex surface, and the carrier 204 may also be formed of a material other than adhesive tape. In other words, the material of the carrier 204 with a convex surface is not particularly limited, as long as the carrier 204 can achieve a convex surface with a protrusion height of 20 μm to 60 μm.

[0083] Fig. 8A FIG. 2 is a cross-sectional view of a carrier 204 having a concave surface according to some embodiments of the present invention. Figure 8B For reference Fig. 8A An exploded top view of the carrier 204. Fig. 8A and Figure 8B In some embodiments, the carrier 204 is an adhesive tape having a concave surface, and the adhesive tape includes a first adhesive layer 204A, a second adhesive layer 204B, and a third adhesive layer 204C. The second adhesive layer 204B is located on the first adhesive layer 204A, wherein the second adhesive layer 204B has a first opening OP1 to expose a portion of the first adhesive layer 204A. The third adhesive layer 204C is located on the second adhesive layer 204B, wherein the third adhesive layer 204C has a second opening OP2 to expose a portion of the second adhesive layer 204B and the first adhesive layer 204A. In addition, the second opening OP2 overlaps the first opening OP1.

[0084] In the embodiment of the present invention, although the carrier 204 with a concave surface is formed by three layers of adhesive tape, the present invention is not limited thereto. In other embodiments, the carrier 204 may also include one or more than three layers of adhesive tape to form the carrier 204 with a concave surface, and the carrier 204 may also be formed of a material other than adhesive tape. In other words, the material of the carrier 204 with a concave surface is not particularly limited, as long as the carrier 204 can achieve a concave depth of 20 μm to 60 μm on the concave surface.

[0085] In some embodiments, the carrier 204 may also be formed by screen printing using ink. The steps of screen printing can refer to Fig. 9 To explain.

[0086] Fig. 9 Schematic cross-sectional view of screen printing in some embodiments of the present invention. Fig. 9As shown, in some embodiments, a dot structure 302 is provided on a substrate 300 to be printed. Among them, the dot structure 302 covers a part of the substrate 300 to be printed and exposes another part of the substrate 300 to be printed. The substrate 300 to be printed can be any substrate or stage for carrying ink, but the present invention is not limited thereto. When screen printing is performed, ink is printed on another part (i.e., the unmasked part) of the substrate 300 through the dot structure 302. After removing the dot structure 302, the ink is vibrated to make the ink slightly creep, and the microscopic ink tends to flatten. The surface ink is partially dissolved with a little solvent to make the microscopic ink more flattened. Finally, it is dried and cured to form a carrier 204 having a concave or convex surface macroscopically.

[0087] In some embodiments, forming a carrier 204 having a concave or convex surface can be achieved by adjusting the design of the dot structure 302. Next, the design of the dot structure 302 will be described.

[0088] FIG. 10A to FIG. 10D Is a top view of a dot structure according to some embodiments of the present invention. Refer to Fig. 10A , in some embodiments, the dot structure 302 includes a plurality of masking portions 302P. When the plurality of masking portions 302P are arranged closer to the center of the dot structure 302, their widths will change from large to small (for example, shrinking from the maximum width W1 to the width W2 and then to the width W3), and their spacings will change from small to large (for example, the spacings SX1, SX2, and SX3 will change from small to large). Accordingly, when screen printing is performed through the dot structure 302 designed in this way, the formed carrier 204 will include the concave surface.

[0089] In Fig. 10A 's embodiments, the masking portions of the dot structure 302 are designed in the form of a dispersed dot structure, but the present invention is not limited thereto. Refer to Fig. 10B , the masking portions 302P of the dot structure 302 are designed to have a continuous arc-shaped structure. As Fig. 10B shown, when the masking portions 302P are arranged closer to the center of the dot structure 302, their widths will change from large to small (W1>W2>W3), and their spacings will change from small to large (SX1<SX2<SX3). Accordingly, when screen printing is performed through the dot structure 302 designed in this way, the formed carrier 204 will include the concave surface.

[0090] Fig. 10A And Fig. 10B 's dot structure 302 can be used to form a carrier 204 having a concave surface. Next, the formation of a carrier 204 having a convex surface will be described.

[0091] Refer to Fig. 10C, in some embodiments, the dot structure 302 includes a plurality of shielding portions 302P. When the plurality of shielding portions 302P are arranged closer to the center of the dot structure 302, their widths increase from small to large (e.g., increasing from the minimum width W3 to the width W2 and then to the width W1), while their spacing decreases from large to small (the closer to the center, the higher the density of the shielding portions 302P). Accordingly, when screen printing is performed through the dot structure 302 designed in this way, the formed carrier 204 will include the convex surface.

[0092] In Fig. 10C the embodiments, the shielding portions of the dot structure 302 are designed in the form of a dispersed dot structure, but the present invention is not limited thereto. Refer to Fig. 10D , the shielding portions 302P of the dot structure 302 are designed to have a continuous arc-shaped structure. As Fig. 10D shown, when the shielding portions 302P are arranged closer to the center of the dot structure 302, their widths increase from small to large (W3 < W2 < W1), while their spacing decreases from large to small (SX3 < SX2 < SX1). Accordingly, when screen printing is performed through the dot structure 302 designed in this way, the formed carrier 204 will include the convex surface.

[0093] In summary, in the method for grinding a silicon carbide wafer of the present invention, when grinding is performed by providing a carrier having a concave surface or a convex surface on the grinding stage, the bent deformation of the wafer after grinding can be offset by stress release, reducing the influence of internal stress release of the wafer on its degree of curvature. Accordingly, a silicon carbide wafer with an ideal flatness can be obtained.

Claims

1. A method for grinding a silicon carbide wafer, characterized in that: include: Cutting the crystal ingot obtained through the crystal growth process to obtain a silicon carbide wafer, wherein the silicon carbide wafer comprises a first surface and a second surface opposite to the first surface; After the silicon carbide wafer is subjected to a grinding and polishing process, the curvature of the silicon carbide wafer is measured; Forming a carrier having a concave surface or a convex surface, wherein the concave surface of the carrier has an inner concave depth of 20 μm to 60 μm, or the convex surface of the carrier has a protrusion height of 20 μm to 60 μm, and the inner concave depth of the concave surface of the carrier or the protrusion height of the convex surface of the carrier is set according to the curvature; Placing the silicon carbide wafer on a grinding stage for a grinding process, so that the carrier having the concave surface or the convex surface is disposed between the grinding stage and the silicon carbide wafer, and the grinding process includes: Placing the silicon carbide wafer on the concave surface or the convex surface of the carrier so that the first surface contacts the concave surface or the convex surface, and performing grinding to remove a portion of the second surface of the silicon carbide wafer; as well as The carrier is removed, and the silicon carbide wafer is placed upside down on the grinding stage, and is ground to remove a portion of the first surface of the silicon carbide wafer.

2. The method according to claim 1, characterized in that When the bearing member has a convex surface, the first surface has residual compressive stress and the second surface has residual tensile stress, and when the bearing member has a concave surface, the first surface has residual tensile stress and the second surface has residual compressive stress.

3. The method according to claim 1, characterized in that: When the carrier has a concave surface, the first surface has a negative curvature, and when the carrier has a convex surface, the first surface has a positive curvature.

4. The method according to claim 1, characterized in that After removing the carrier, the silicon carbide wafer is reversely placed on the flat surface of the second carrier on the grinding stage so that the second surface contacts the flat surface, and then ground to remove a portion of the first surface of the silicon carbide wafer.

5. The method according to claim 1, characterized in that The carrier is an adhesive tape including the concave surface, and the adhesive tape includes: First adhesive layer; a second adhesive layer, located on the first adhesive layer, wherein the second adhesive layer has a first opening to expose a portion of the first adhesive layer; and The third glue layer is located on the second glue layer, wherein the third glue layer has a second opening to expose the second glue layer and a portion of the first glue layer, and the second opening overlaps the first opening.

6. The method according to claim 1, characterized in that The carrier is an adhesive tape including the convex surface, and the adhesive tape includes: A first adhesive layer having a first diameter DA1; a second adhesive layer, located on the first adhesive layer and having a second diameter DA2; and The third adhesive layer is located on the second adhesive layer and has a third diameter DA3, wherein DA1>DA2>DA3.

7. The method according to claim 1, characterized in that The carrier is formed by screen printing using ink, and the steps of screen printing include: Providing a dot structure on a printed object, wherein the dot structure masks a portion of the printed object and exposes another portion of the printed object; Performing screen printing so that ink is printed on the other part of the printed object through the halftone structure; and The steps of vibrating and drying are performed to form the carrier.

8. The method according to claim 7, characterized in that The mesh dot structure includes a plurality of shielding parts, and when the plurality of shielding parts are arranged closer to the center of the mesh dot structure, their widths will change from large to small and their spacings will change from small to large, and the formed carrier includes the convex surface.

9. The method according to claim 7, characterized in that: The mesh dot structure includes a plurality of shielding parts, and when the plurality of shielding parts are arranged closer to the center of the mesh dot structure, their widths will change from small to large and their spacings will change from large to small, and the formed carrier includes the concave surface.

10. A silicon carbide wafer manufactured by the method according to claim 1, characterized in that: The curvature of the silicon carbide wafer after the grinding process is within a range of +25 μm.

11. A method for grinding a silicon carbide wafer, characterized in that: include: Cutting the crystal ingot obtained through the crystal growth process to obtain a silicon carbide wafer, wherein the silicon carbide wafer comprises a first surface and a second surface opposite to the first surface, the first surface has residual tensile stress and the second surface has residual compressive stress; Placing the silicon carbide wafer on a grinding stage for a grinding process, wherein a carrier having a concave surface is provided between the grinding stage and the silicon carbide wafer, and the grinding process comprises: Placing the silicon carbide wafer on the concave surface of the carrier so that the first surface contacts the concave surface, and performing grinding to remove a portion of the second surface of the silicon carbide wafer; as well as The carrier is removed, and the silicon carbide wafer is placed upside down on the grinding stage, and is ground to remove a portion of the first surface of the silicon carbide wafer.

Images