Search

CN-122008041-A - Wafer surface low-damage efficient repairing method based on grinding wheel optimization

CN122008041ACN 122008041 ACN122008041 ACN 122008041ACN-122008041-A

Abstract

The invention discloses a wafer surface low-damage efficient repairing method based on grinding wheel optimization, and relates to the technical field of wafer regeneration processing. According to the invention, the design that the grinding wheels with multiple meshes are arranged in series and are polished in a rough-to-fine gradient manner is adopted, so that the complicated processes of reciprocating polishing and frequent mesh switching of a single grinding wheel in the prior art are omitted, the seamless connection of working procedures is realized, and the processing efficiency is greatly improved. Meanwhile, the real-time water injection cooling of the grinding wheel in the grinding process is realized through the matching of the liquid storage cavity and the liquid leakage port, the processing temperature is effectively controlled, and the thermal deformation and subsurface damage of the wafer caused by local overheating in the prior art are avoided.

Inventors

  • GAO XIANG

Assignees

  • 安徽富乐德长江半导体材料股份有限公司

Dates

Publication Date
20260512
Application Date
20260326

Claims (8)

  1. 1. The wafer surface low-damage high-efficiency repairing method based on grinding wheel optimization is characterized by comprising the following steps of: s1, placing a wafer, namely placing a wafer body (400) at the center of the top end of a placing ring (205); S2, carrying out multistage polishing, namely rotationally polishing the wafer body (400) by adopting grinding wheels with different meshes, wherein the meshes of the grinding wheels used for polishing are sequentially from low to high; S3, stability verification, namely continuously processing a plurality of wafers containing defects, and verifying the stability of the heavy scratch removal rate, TTV and subsurface damage layer depth index.
  2. 2. The method for efficiently repairing the low-damage surface of the wafer based on grinding wheel optimization of claim 1 is characterized in that in the step S1, a plurality of supporting rods (204) are uniformly and fixedly connected to the bottom end of a placing ring (205), a bonding disc (201) is fixedly connected to the bottom end of each supporting rod (204), a vacuum pump (206) is fixedly connected to the center of the top end of each bonding disc (201), an adsorption disc (207) is fixedly connected to the top end of each vacuum pump (206), a plurality of vacuum holes (208) are formed in the top end of each adsorption disc (207) in a penetrating mode, the adsorption discs (207) are adsorbed on the back surface of the wafer body (400) through atmospheric pressure difference to form a negative pressure area, and the vacuum holes (208) are used for avoiding deformation of the wafer body (400) caused by overlarge local pressure.
  3. 3. The method for efficiently repairing the low-damage surface of the wafer based on grinding wheel optimization is characterized in that a rotating ring (202) is fixedly connected to the side face of a bonding disc (201), a base plate (101) is arranged at the bottom end of the bonding disc (201), a positioning groove (102) is formed in the position, below the bonding disc (201), of the top end of the base plate (101), a rotating groove (103) is formed in the side face of the positioning groove (102), the rotating ring (202) is rotationally connected to the inner portion of the rotating groove (103), a motor groove (104) is formed in the center of the bottom end of the positioning groove (102), a rotating motor I (203) is fixedly connected to the inner portion of the motor groove (104), the bottom end of the bonding disc (201) is fixedly connected to the output end of the rotating motor I (203), and the rotating motor I (203) is used for driving the wafer body (400) subjected to adsorption and polishing to rotate in the opposite direction.
  4. 4. The method for efficiently repairing the low-damage surface of the wafer based on the grinding wheel optimization of claim 3, wherein in the step S2, a telescopic motor (301) is fixedly connected to the center of the top end of the chassis (101), a telescopic rod (302) is fixedly connected to the output end of the top end of the telescopic motor (301), a serial disc (303) is rotatably connected to the top end of the telescopic rod (302), a rotating motor II (304) is arranged at the bottom end of the serial disc (303), the output end of the rotating motor II (304) is fixedly connected to the top end of the serial disc (303), and a plurality of L-shaped connecting rods (305) are fixedly connected to the side face of the serial disc (303).
  5. 5. The method for efficiently repairing the surface of the wafer with low damage based on grinding wheel optimization of claim 4, wherein one end, far away from the serial disc (303), of the L-shaped connecting rod (305) is fixedly connected with a third rotating motor (306), the output end at the bottom end of the third rotating motor (306) is fixedly connected with a liquid storage cavity (307), and the rotating directions of the third rotating motor (306) and the first rotating motor (203) are opposite.
  6. 6. The method for efficiently repairing the surface of the wafer with low damage based on grinding wheel optimization of claim 5, wherein the outer ring of the bottom end of the liquid storage cavity (307) is provided with threaded grooves (310), threaded rings (311) are connected in the threaded grooves (310) in a threaded manner, thickness rings (313) are fixedly connected to the bottom ends of the threaded rings (311), grinding wheel placing grooves (312) are formed between the thickness rings (313) and the threaded rings (311), and grinding wheels with different numbers are placed in sequence according to the order of the numbers of the grinding wheels in the grinding wheel placing grooves (312).
  7. 7. The method for efficiently repairing the surface of the wafer based on grinding wheel optimization of claim 6, wherein the top end of the liquid storage cavity (307) is movably connected with a sealing plug (308), the bottom end of the liquid storage cavity (307) is provided with a liquid leakage port (309), the sealing plug (308) and the liquid leakage port (309) are used for injecting water to the surface of the grinding wheel when the wafer body (400) is rotationally polished, and the water injection is used for avoiding deformation of the wafer body (400) caused by local overheating.
  8. 8. The method for efficiently repairing the low damage of the wafer surface based on the grinding wheel optimization of claim 1, wherein in the step S3, the wafer body (400) used for verification is a 12-inch monocrystalline Si wafer, a step of rapidly detecting the surface of the wafer body (400) is arranged at a process joint part between each stage of multistage polishing, an online laser thickness meter is used for detecting whether the removal amount of the wafer body (400) reaches a polishing target or not in the step of detecting, and a preliminary removal effect of heavy scratches is rapidly judged through a high-speed vision camera.

Description

Wafer surface low-damage efficient repairing method based on grinding wheel optimization Technical Field The invention relates to the technical field of wafer regeneration processing, in particular to a wafer surface low-damage high-efficiency repairing method based on grinding wheel optimization. Background In the subsequent steps of semiconductor wafer regeneration and chip manufacturing, thinning the back of the wafer is a key step of the subsequent steps of wafer regeneration and chip manufacturing, and is a key for improving the integration level and heat dissipation performance of devices, the main stream in the industry adopts a rough grinding and fine grinding double-shaft grinding process, materials are rapidly removed through a low-mesh grinding wheel, macroscopic defects are eliminated, and a damaged layer is removed through a high-mesh grinding wheel, so that a smooth surface is obtained. However, when the existing conventional process is used for processing a heavily scratched wafer, the core contradiction of deep rough grinding damage layer and low accurate grinding efficiency is faced, so that the overall processing efficiency is low, the material removal amount and cost of a downstream polishing process are increased due to the residual damage layer, meanwhile, the severe requirements of advanced packaging on low damage and high flatness of the wafer are met, the damage control capability of the conventional grinding process is more challenging, and the balance of efficient repair of defects, accurate damage control and optimization of the whole process chain cost is needed to be realized through process optimization. Insufficient finish grinding can leave a damaged layer, so that the material removal amount is forced to be increased in the subsequent polishing process, and further, the consumption of polishing liquid is increased, the working time is prolonged and the cost is increased. Related patents related to wafer surface defect repair relate to the following: The application discloses a three-dimensional morphology repair process for heavy scratching of a wafer based on a DFG8560 single-axis fine grinding mode, particularly discloses a three-dimensional morphology repair process for heavy scratching of a wafer based on a DFG8560 single-axis fine grinding mode, and aims to solve the problems of deep subsurface damage, low processing efficiency and serious material waste caused by rough grinding in a traditional double-stage grinding process. The process comprises the steps of identifying heavy scratch defects on the surface of a wafer through an automatic optical detection system, adopting a diamond grinding wheel with 8000 ceramic bond particle sizes, carrying out single-axis fine grinding through ultra-pure water cooling under the optimized parameters of 1300-2300rpm of a main shaft rotating speed, 1785-2100rpm of a workbench rotating speed and 0.6-1.3 of a rotating speed ratio, skipping rough grinding to directly finish three-dimensional morphology repair, and dynamically regulating and controlling the rotating speed and the pressure in three stages, wherein the total time is less than or equal to 135 seconds, and the removal amount is 6-12 microns. By adopting the technical scheme, the application can realize the surface roughness Ra of <0.5 nanometers, the subsurface damage layer of <50 nanometers and the repair success rate of >98.5 percent, obviously shorten the processing time by more than 35 percent and reduce the material removal amount by 40 percent. However, in the prior patent, the wafer can be repaired with high efficiency, low damage and high continuity by heavily scratching the three-dimensional shape. But the process skips rough grinding and directly adopts 8000-mesh grinding wheels to uniaxially fine grind, so that the high-mesh grinding wheels have limited cutting capability for heavy scratched wafers with larger depth, and the conditions of insufficient grinding force and incomplete defect removal are easy to occur, so that the suitability for heavy scratching is limited. The single-shaft fine grinding only depends on a single grinding wheel to finish all repairing work, the abrasive particle abrasion speed of the grinding wheel can be greatly accelerated, the service life of the grinding wheel is shortened, the replacement cost of consumable materials is increased, the repairing precision of later processing is easy to fluctuate due to abrasion of the grinding wheel, and the mass production stability is influenced. Disclosure of Invention The application aims to provide a wafer surface low-damage high-efficiency repairing method based on grinding wheel optimization, which is used for solving the problem of low processing efficiency in the process of polishing and repairing a wafer in the prior art. The invention provides a method for repairing heavy scratches on the surface of a wafer with high efficiency and low damage by optimizing different combinations of the m