Silicon carbide crystal cutting system and method based on 3D defect detection

Abstract

The invention relates to a silicon carbide crystal cutting system and method based on 3D defect detection, and the silicon carbide crystal cutting system based on 3D defect detection comprises an X-ray CT scanning device which obtains the internal structure and defect information of a crystal; the silicon carbide crystal fixing equipment is used for fixing silicon carbide crystals; the 3D data processing module is connected with the X-ray CT scanning equipment and comprises a defect data module, a defect data comparison and analysis module and an image processing and generating module; the cutting data processing module is used for receiving 3D defect data and images and then processing and calculating the 3D defect data and images to form an optimal cutting path method; the cutting system is connected with the cutting data processing module, cutting process parameters are set according to the optimal cutting path, and finally a silicon carbide wafer is formed; according to the method, the optimal cutting route is obtained through 3D defect scanning and defect data analysis, the low-defect high-yield wafer is formed, the method adapts to different defect distributions, the cutting precision and efficiency are improved, and the machining loss is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of silicon carbide cutting, and specifically relates to a silicon carbide crystal cutting system and a cutting method based on 3D defect detection. Background Art

[0002] Since the mid-twentieth century, silicon materials, as semiconductor materials, have been widely used in the field of integrated circuit chips in low-voltage, low-frequency, low-power transistors and detectors, including computers, mobile phone CPUs, GPUs, system-on-chips (SoCs), etc. As third-generation semiconductor materials, silicon carbide (SiC) and gallium nitride (GaN), especially SiC, have significant characteristics such as high temperature resistance and high voltage resistance. They are currently widely used in rail transportation, new energy vehicles, wind power and other fields. From a process point of view, the silicon carbide production process mainly includes: ingot growth, slicing, grinding and polishing, epitaxy, lithography, etching, ion implantation, deposition and other steps. Because silicon carbide is one of the hardest substances, up to Mohs 9.2 and 9.3, in the silicon carbide chip manufacturing process, iterative ingot cutting technology and slicing technology are currently the main ways to increase production capacity. Traditional multi-wire cutting technology is a process for cutting brittle and hard materials; mortar wire cutting technology has been applied to most silicon carbide manufacturers, and diamond wire cutting technology has also become a mainstream iterative solution. The above methods not only have low cutting efficiency and high material loss, but also high consumables cost. The latest laser cutting technology has high processing efficiency and low material loss, and is currently a research hotspot for cutting methods in the industry.

[0003] Patent publication number CN118081117A discloses a novel silicon carbide cutting system and method, which uses femtosecond laser equipment for cutting. Since femtosecond laser has the advantages of extremely short pulse width, extremely high peak power, low heat input, and wide adjustable depth-to-diameter ratio, when cutting silicon carbide wafers and ingots, gratings and galvanometers can be used to output multiple beams for simultaneous cutting for wafer cutting, and a large depth-to-diameter ratio can be adjusted for ingot cutting, and multiple iterative cutting can be performed, while secondary cutting can be performed with the aid of plasma etcher shaping, cleaning, and visual and infrared real-time monitoring. Cutting, cutting efficiency is greatly improved, and material loss is reduced; Patent Publication No. CN117601293A discloses a method for cutting silicon carbide, which uses a heater in conjunction with a high-pressure water column to reduce the cost of hardware purchase and effectively cut silicon carbide; CN117283152A discloses a method for high-efficiency laser cutting of silicon carbide ingots, which adopts a dual-wavelength laser coaxial synchronous single processing mode to achieve the simultaneous and spatial action of the modified laser and the heating laser on the SiC ingot, reducing the processing time, the difficulty of wafer stripping, and improving the stripping quality.

[0004] Patent publication number CN116465912A discloses a method for rapid nondestructive testing of silicon carbide crystal quality. The processed silicon carbide crystal is fully scanned through a specific X-ray light source and scanning method. Based on the local differences in X-ray absorption or scattering by different defects in the crystal material, the defect density of the entire or local silicon carbide crystal can be quickly obtained on a computer tomography scan, as well as an intuitive image of the distribution.

[0005] In summary: Currently, silicon carbide crystal cutting mainly improves cutting efficiency and reduces cutting losses by changing cutting tools and improving cutting methods. At the same time, the detection accuracy of silicon carbide crystals still needs to be further improved. Summary of the invention

[0006] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: a silicon carbide crystal cutting system based on 3D defect detection, comprising:

[0008] X-ray tomography (CT) scanning equipment, used to perform all-round, high-resolution 3D scanning of silicon carbide crystals to obtain the structure and defect information inside the crystals;

[0009] Silicon carbide crystal fixing equipment, used to fix silicon carbide crystals, with the functions of rotation and vertical movement up and down, to ensure that the position of the crystal is accurately controlled during scanning and cutting;

[0010] The 3D data processing module is connected to the X-ray tomography CT scanning device, and includes a defect data module, a defect data comparison and analysis module, and an image processing and generation module. After the 3D data processing module obtains the X-ray data signal, it obtains a 3D defect image of the silicon carbide crystal.

[0011] The cutting data processing module is connected to the 3D data module. After receiving the 3D defect data and images, the cutting data module processes and calculates to form the best cutting path method;

[0012] The cutting system is connected to the cutting data processing module, sets the cutting process parameters according to the optimal cutting path, and finally forms the silicon carbide wafer.

[0013] As a preferred technical solution for a silicon carbide crystal cutting system based on 3D defect detection, the X-ray tomography CT scanning equipment is a micro-focus high-energy X-ray tube, the acceleration voltage range of the X-ray light source is 200-350kV, the focal size of the light source is ≤10 microns, the exposure time is ≥1000ms, and the spatial resolution is ≤80 microns.

[0014] As a preferred technical solution for a silicon carbide crystal cutting system based on 3D defect detection, the defect data module includes a micropipe defect data module, a carbon wrapping defect data module, and a polycrystalline / polytype defect data module; the defect data comparison and analysis module classifies different defect data; the image generation module assigns different defect types different colors based on the comparison and analysis results, and data within the normal range is represented by black.

[0015] As a preferred technical solution for a silicon carbide crystal cutting system based on 3D defect detection, based on 3D defect image data, the cutting data processing module includes a defect distribution analysis module and a cutting path analysis module. The cutting path analysis module of the cutting data processing module is provided with 4 silicon carbide defect level divisions, and the Class A defect index is <1000 / cm 2 、Class B level defect index 1000-3000 / cm 2 ; Class C defect index 3000-8000 / cm 2 ; Class D defect index >8000 / cm 2 .

[0016] The present invention also provides a cutting method based on the aforementioned silicon carbide crystal cutting system based on 3D defect detection, comprising:

[0017] Step 1: Clean the surface of the silicon carbide crystal;

[0018] Step 2: Place the silicon carbide crystal on a fixed device, set the rotation or movement speed, and the X-ray tomography CT scan emits X-rays, and the silicon carbide crystal rotates or moves up and down to obtain X-ray data signals;

[0019] Step 3: The X-ray data signal is input into the 3D data processing module, and the X-ray data is processed and analyzed according to the difference in X-ray absorption or scattering of different defect parts, and finally a 3D defect image is formed through the image generation module;

[0020] Step 4: The 3D defect image data is input into the cutting data module, which performs data analysis based on the four silicon carbide defect levels to obtain the best cutting path;

[0021] Step 5: The cutting system performs cutting according to the optimal cutting path and selects appropriate cutting parameters to ultimately obtain silicon carbide wafers of different specifications.

[0022] As a preferred technical solution for a silicon carbide crystal cutting method based on 3D defect detection, in step 4, the cutting data module performs data analysis based on four silicon carbide defect levels, and sets the cutting path to pass through the defect-dense area to the greatest extent, and the first cutting finally obtains silicon carbide wafers of different thicknesses with minimized surface defects;

[0023] For silicon carbide wafers of different thicknesses with minimized surface defects obtained by the first cutting, a cutting path analysis can be further performed, and a second cutting can be performed to obtain silicon carbide wafers with minimized defects and meeting thickness requirements.

[0024] As a preferred technical solution for a silicon carbide crystal cutting method based on 3D defect detection, in step 4, the cutting data module performs data analysis based on four silicon carbide defect levels, and sets the thickness of the final cut silicon carbide wafer to be consistent, with a thickness of 0.35 mm or 0.5 mm, and finally obtains the optimal cutting path with consistent wafer thickness.

[0025] As a preferred technical solution for a silicon carbide crystal cutting method based on 3D defect detection, the X-ray data signal is input into a 3D data processing module, specifically including:

[0026] The X-ray data is processed using the FDK cone-beam CT reconstruction algorithm to reconstruct a 3D model of the silicon carbide crystal, where the reconstructed voxel intensity I(x, y, z) represents the attenuation coefficient of each point in the crystal;

[0027] Artificial intelligence algorithms such as deep learning neural networks trained with labeled defect data are used to analyze the 3D model and identify defect types such as microtubes, carbon packaging, and polycrystalline areas. The neural network outputs the defect classification probability P for each voxel. defect (x,y,z).

[0028] As a preferred technical solution for a silicon carbide crystal cutting method based on 3D defect detection, 3D defect image data is input into a cutting data processing module, including:

[0029] Use machine learning algorithms trained with historical defect and cutting data to analyze the spatial distribution of defects and predict the best cutting plane;

[0030] The optimal cutting path is calculated by minimizing the total defect density in the resulting wafer, using the formula:

[0031] Minimize D 总 =∑ i (D i ×h i )

[0032] Among them, D 总 is the total defect density, Di is the defect density of the i-th layer, t i is the thickness of the i-th layer;

[0033] Cutting parameters for each cutting path segment, including cutting depth, speed, and power settings, are determined based on defect data and material properties.

[0034] As a preferred technical solution for a silicon carbide crystal cutting method based on 3D defect detection, cutting is performed according to an optimal cutting path, including:

[0035] Inputting cutting parameters from the cutting data processing module into the cutting system;

[0036] Start the cutting process, the laser cuts according to the predetermined path, and the initial settings are a cutting line width of 0.15mm, a cutting speed of 3mm / s, and a laser peak power of 1.5GW;

[0037] The cutting process is continuously monitored, and cutting parameters such as laser power, focus and speed are adjusted in real time based on the feedback from the 3D data processing module and the in-situ sensor to cope with the changes in defect distribution encountered during the cutting process.

[0038] Beneficial effects of the present invention: The present invention adjusts parameters such as laser power and cutting speed in real time to adapt to different defect distributions, thereby improving cutting accuracy and efficiency and reducing processing losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0040] Figure 1 It is a schematic diagram of the working process of the cutting system of the present invention;

[0041] Figure 2 This is a schematic diagram of the cutting route entry point of the present invention. DETAILED DESCRIPTION

[0042] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0043] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0044] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0045] Secondly, the present invention is described in detail with reference to the schematic diagram. When describing the embodiments of the present invention in detail, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0046] Example 1

[0047] Reference Figure 1 and 2 , this embodiment provides a silicon carbide crystal cutting system based on 3D defect detection, including:

[0048] X-ray tomography (CT) scanning equipment, used to perform all-round, high-resolution 3D scanning of silicon carbide crystals to obtain the structure and defect information inside the crystals;

[0049] Silicon carbide crystal fixing equipment, used to fix silicon carbide crystals, with the functions of rotation and vertical movement up and down, to ensure that the position of the crystal is accurately controlled during scanning and cutting;

[0050] A 3D data processing module, connected to the X-ray tomography CT scanning device, includes a defect data module for collecting and storing data on micropipe defects, carbon encapsulation defects, and polycrystalline / polymorphic defects;

[0051] That is, the raw data obtained from X-ray CT scans are collected and stored.

[0052] Defect data comparison and analysis module, which uses artificial intelligence algorithms to classify and analyze defect data;

[0053] That is, preprocessing, segmenting and classifying the raw data to identify different types of defects;

[0054] The image processing and generation module generates 3D defect images with different colors representing different defect types.

[0055] The cutting data processing module is connected with the 3D data processing module and includes:

[0056] The defect distribution analysis module uses machine learning algorithms to analyze the spatial distribution of defects. It uses the trained convolutional neural network (CNN) model to conduct deep learning on defect data and predict the optimal cutting path.

[0057] Cutting path optimization module that uses deep learning algorithms to calculate the best cutting path to maximize the yield of low-defect wafers.

[0058] The cutting system, connected to the cutting data processing module, is capable of performing cutting according to the optimal cutting path and adjusting cutting parameters in real time based on feedback from the 3D data processing module.

[0059] The relationship between the modules is:

[0060] The X-ray tomography (CT) scanning device acquires the internal structure data of the silicon carbide crystal, and the data is transmitted to the 3D data processing module.

[0061] The 3D data processing module processes the data to generate a 3D defect image, calculates the best cutting path through the artificial intelligence optimization module, and sends the result to the cutting data processing module.

[0062] The cutting data processing module generates specific cutting parameters based on the optimal cutting path and transmits them to the cutting system.

[0063] The cutting system executes cutting according to the parameters and returns real-time feedback information (such as cutting progress and abnormal conditions) to the cutting data processing module to achieve closed-loop control.

[0064] Furthermore, the X-ray tomography CT scanning device is a micro-focus high-energy X-ray tube, the acceleration voltage range of the X-ray light source is 200-350 kV, the focal spot size of the light source is ≤10 microns, the exposure time is ≥1000 ms, and the spatial resolution is ≤80 microns.

[0065] Furthermore, the 3D data processing module uses a deep learning algorithm to classify defects with higher accuracy.

[0066] Furthermore, the cutting data processing module optimizes the cutting path using a neural network trained with historical defect data and cutting results.

[0067] Furthermore, the cutting system is a laser cutting system capable of adjusting cutting parameters in real time according to feedback from the 3D data processing module.

[0068] The defect data module includes the micro-pipe defect data module, the carbon package defect data module, and the polycrystalline / polytype defect data module; the defect data comparison and analysis module classifies different defect data; the image generation module assigns different defect types different colors based on the comparison and analysis results, and the data within the normal range is represented by black.

[0069] Based on 3D defect image data, the cutting data processing module contains a defect distribution analysis module and a cutting path analysis module. The cutting path analysis module of the cutting data processing module is provided with 4 silicon carbide defect level divisions, and the Class A defect index is <1000 / cm 2 、Class B level defect index 1000-3000 / cm 2 ; Class C defect index 3000-8000 / cm 2 ; Class D defect index >8000 / cm 2 ;

[0070] By implementing the following technical steps, the present invention realizes efficient and high-quality cutting of silicon carbide crystals, and the specific analysis is as follows:

[0071] Use of X-ray CT scanning equipment:

[0072] By introducing high-energy micro-focus X-ray CT scanning equipment into the system, a full-scale, high-resolution 3D scan of silicon carbide crystals is achieved, and the structure and defect information inside the crystal is obtained. By using an X-ray tube with an acceleration voltage range of 200-350kV and a focus size of ≤10 microns, the depth and accuracy of the scan are guaranteed. This step makes it possible to accurately understand the defect distribution inside the crystal before cutting, including microtubes, carbon packaging, polycrystalline areas, etc., providing an accurate data basis for subsequent defect identification and cutting path optimization. Non-destructive detection of internal defects in the crystal is achieved, avoiding cutting quality problems caused by the lack of internal defect information in traditional methods, and significantly improving the yield rate of wafers.

[0073] Design of silicon carbide crystal fixing equipment:

[0074] By using a silicon carbide crystal fixing device with the functions of rotation and vertical movement, the crystal can be precisely positioned and controlled to move during scanning and cutting. By precisely controlling the position and angle of the crystal, the X-ray CT scanning device can obtain information inside the crystal from multiple angles and positions. This step ensures the integrity and accuracy of the scan data, and at the same time, during the cutting process, ensures the consistency of the cutting path with the optimization plan, avoids cutting deviations caused by positioning errors, and improves cutting accuracy and efficiency.

[0075] Application of 3D data processing module:

[0076] Use of the Defect Data Module

[0077] By collecting and storing data on micropipe defects, carbon encapsulation defects, and polycrystalline / polymorphic defects, systematic management and analysis of different types of defects are achieved, providing comprehensive data support for subsequent defect classification and cutting path optimization.

[0078] Artificial intelligence algorithm of defect data comparison and analysis module:

[0079] Artificial intelligence algorithms such as deep learning neural networks are used to classify and analyze defect data. By training a model with labeled defect data, different defect types in the crystal can be accurately identified. The neural network outputs the defect classification probability P for each voxel. defect (x,y,z). It realizes high-precision identification and classification of defects, and improves the accuracy and efficiency of defect identification.

[0080] Visualization of image processing and generation modules:

[0081] Generate 3D defect images with different colors representing different defect types, and visually display the spatial distribution of defects. This step helps technicians quickly understand the defects inside the crystal, assists in decision-making, and improves the operability of the system.

[0082] Functions of cutting data processing module:

[0083] Machine learning algorithm for defect distribution analysis module:

[0084] The convolutional neural network (CNN) trained with historical defects and cutting data is used to analyze the spatial distribution of defects and predict the optimal cutting plane. Through the deep learning algorithm, the optimal cutting path is calculated by comprehensively considering factors such as defect type, density and location. The intelligent and optimized cutting plan is realized, maximizing the output rate of low-defect wafers.

[0085] Deep learning algorithm of cutting path optimization module:

[0086] The optimal cutting path is calculated using a deep learning algorithm with the goal of minimizing the total defect density in the resulting wafer, using the formula:

[0087] Minimize D 总 =∑ i (D i ×h i )

[0088] Among them, D 总 is the total defect density, D i is the defect density of the i-th layer, h i is the thickness of the i-th layer. Through accurate calculation of mathematical models, the cutting path is optimized and the wafer quality is improved.

[0089] Real-time feedback and adjustment of cutting system:

[0090] The cutting system is able to perform cutting according to the optimal cutting path and adjust the cutting parameters in real time based on the feedback from the 3D data processing module. During the cutting process, the cutting status is continuously monitored, and parameters such as laser power, focus position and cutting speed are adjusted in real time according to the changes in defect distribution. Closed-loop control is achieved, which improves the stability and reliability of the cutting process and ensures the quality of the final wafer.

[0091] Adoption of high-performance X-ray CT equipment:

[0092] By using a micro-focus high-energy X-ray tube with an acceleration voltage range of 200-350kV, a focal spot size ≤10 microns, an exposure time ≥1000ms, and a spatial resolution ≤80 microns, the depth and resolution of the scan are improved. It is possible to detect tiny defects, improve the accuracy of defect identification, and provide high-quality data support for subsequent cutting optimization.

[0093] Application of Deep Learning Algorithms:

[0094] The 3D data processing module uses a deep learning algorithm to classify defects with higher accuracy. Through deep learning, the ability to identify complex defects is improved, misjudgment is reduced, and the intelligence level of the system is improved.

[0095] Cutting path optimization based on historical data:

[0096] The cutting data processing module uses a neural network trained with historical defect data and cutting results to optimize the cutting path. Using historical data, the model can more accurately predict the best cutting solution, reducing trial and error costs and improving cutting efficiency and wafer quality.

[0097] Real-time adjustment laser cutting system: The cutting system is a laser cutting system that can adjust cutting parameters in real time according to the feedback from the 3D data processing module. By adjusting parameters such as laser power and cutting speed in real time, it can adapt to different defect distributions, improve cutting accuracy and efficiency, and reduce processing losses.

[0098] Example 2

[0099] The present invention also discloses a cutting method of a silicon carbide crystal cutting system based on 3D defect detection, comprising step 1: performing surface cleaning treatment on the silicon carbide crystal;

[0100] Step 2: Place the silicon carbide crystal on a fixed device, set the rotation or movement speed, and the X-ray tomography CT scan emits X-rays, and the silicon carbide crystal rotates or moves up and down to obtain X-ray data signals;

[0101] Step 3: The X-ray data signal is input into the 3D data processing module, and the X-ray data is processed and analyzed according to the difference in X-ray absorption or scattering of different defect parts, and finally a 3D defect image is formed through the image generation module;

[0102] Step 4: The 3D defect image data is input into the cutting data module, which performs data analysis based on the four silicon carbide defect levels to obtain the best cutting path;

[0103] Step 5: The cutting system performs cutting according to the optimal cutting path and selects appropriate cutting parameters to ultimately obtain silicon carbide wafers of different specifications.

[0104] Furthermore, the cutting data module performs data analysis based on four SiC defect levels and sets the cutting path to pass through the defect-dense area to the greatest extent possible. The first cutting ultimately obtains SiC wafers of different thicknesses with minimized surface defects.

[0105] For the silicon carbide wafers of different thicknesses with minimized surface defects obtained by the first cutting, further cutting path analysis can be performed, and the silicon carbide wafers with minimized defects and meeting thickness requirements can be obtained by the second cutting;

[0106] The cutting data module performs data analysis based on four silicon carbide defect levels and sets the thickness of the final cut silicon carbide wafers to be consistent, at 0.35mm or 0.5mm, ultimately obtaining the optimal cutting path with consistent wafer thickness.

[0107] Furthermore, in step 4, the cutting data module performs data analysis based on the four silicon carbide defect levels, and sets the thickness of the silicon carbide wafers to be finally cut to be inconsistent, with thicknesses of 0.35 mm and 0.5 mm, and finally obtains the optimal cutting path for inconsistent wafer thickness.

[0108] The cutting method can be specifically:

[0109] Step (1) cleaning the surface of the silicon carbide crystal with ethanol or isopropanol to remove contaminants, and then drying with nitrogen to prevent oxidation;

[0110] Step (2) fixing the silicon carbide crystal on a crystal fixing device with the function of rotating and moving up and down, setting the rotation speed to 1 revolution per minute, starting the X-ray tomography CT scanning device, emitting X-rays while the crystal rotates, and acquiring X-ray data signals from multiple angles;

[0111] Step (3) inputting the X-ray data signal into the 3D data processing module specifically includes:

[0112] The X-ray data were processed using the Feldkamp-Davis-Kress (FDK) cone-beam CT reconstruction algorithm to reconstruct a 3D model of the silicon carbide crystal, where the reconstructed voxel intensity I (x, y, z) represents the attenuation coefficient of each point in the crystal;

[0113] Artificial intelligence algorithms such as deep learning neural networks trained with labeled defect data are used to analyze the 3D model and identify defect types such as microtubes, carbon packaging, and polycrystalline areas. The neural network outputs the defect classification probability P for each voxel. defect (x,y,z);

[0114] The defect density D of different areas is calculated using the formula D = N / A, where N is the number of defects detected in area A;

[0115] Defects are classified into four predefined levels based on the defect density per unit area:

[0116] Grade A: D<1000 pieces / cm 2 ;

[0117] Grade B: D is 1000-3000 pieces / cm 2 between;

[0118] C grade: D is 3000-8000 pieces / cm 2 between;

[0119] D grade: D>8000 pieces / cm 2 ;

[0120] Generate 3D defect images, with different defect types and levels represented by different colors for easy visualization;

[0121] Step (4) inputting the 3D defect image data into the cutting data processing module specifically includes:

[0122] Use machine learning algorithms (such as convolutional neural networks (CNNs)) trained with historical defect and cutting data to analyze the spatial distribution of defects and predict the best cutting plane;

[0123] The optimal cutting path is calculated by minimizing the total defect density in the resulting wafer, using the formula:

[0124] Minimize D 总 =∑ i (D i ×h i )

[0125] Among them, D 总 is the total defect density, D i is the defect density of the i-th layer, h i is the thickness of the i-th layer;

[0126] determining cutting parameters for each cutting path segment based on the defect data and material properties, including cutting depth, speed, and power settings;

[0127] Step (5) setting up a cutting system (a laser cutting system capable of adjusting parameters in real time) and performing cutting according to an optimal cutting path, specifically comprising:

[0128] Inputting cutting parameters from the cutting data processing module into the cutting system;

[0129] Start the cutting process, the laser cuts according to the predetermined path, and the initial settings are a cutting line width of 0.15mm, a cutting speed of 3mm / s, and a laser peak power of 1.5GW;

[0130] Continuously monitor the cutting process and adjust cutting parameters such as laser power, focus and speed in real time based on feedback from the 3D data processing module and in-situ sensors to cope with changes in defect distribution encountered during the cutting process;

[0131] Step (6) collects the cut silicon carbide wafers. Due to the optimized cutting path and real-time parameter adjustment, the wafer defects are minimized, meeting or exceeding the market thickness (such as 0.35 mm) and quality requirements, and reaching the A-level defect level.

[0132] In step (3), the 3D data processing module classifies the defects into four levels according to the defect density and represents them with different colors.

[0133] In step (4), the cutting data processing module uses the defect classification to set the cutting path so that it passes through the defect-intensive area, thereby minimizing the defects in the resulting wafer.

[0134] In step (5), the cutting system adjusts parameters such as cutting speed and power in real time according to the feedback from the 3D data processing module to ensure the best cutting quality.

[0135] After the initial dicing, the diced wafers undergo secondary 3D defect inspection and analysis and, if necessary, a second dicing is performed to further minimize defects and meet market thickness requirements.

[0136] Furthermore, in step (3), X-ray data signal processing and defect identification are required.

[0137] In this step, the data signal obtained by the X-ray tomography CT scanning device enters the 3D data processing module. The specific process is as follows:

[0138] Data preprocessing:

[0139] The two-dimensional cross-sectional image sequence obtained by X-ray CT scanning is reconstructed to form three-dimensional volume data.

[0140] Perform denoising on the data, such as using Gaussian filtering or median filtering, to eliminate the noise generated during the scanning process.

[0141] Defect type identification:

[0142] Voxel classification: Divide the three-dimensional volume data into a number of voxels, each of which represents a tiny cubic area in the crystal.

[0143] Feature extraction: For each voxel, calculate its grayscale value, texture features and other parameters.

[0144] Neural Network Classification:

[0145] Model Input: The features of each voxel are input into a pre-trained convolutional neural network (CNN) model.

[0146] Model output: The CNN model outputs the probability value of each voxel belonging to a specific defect type, denoted as Where i, j, k are the positions of the voxels in three-dimensional space, and C is the defect type (such as microtubes, carbon inclusions, polycrystalline regions, etc.).

[0147] Determination of defect type:

[0148] For each voxel, the defect type with the highest probability is selected as its classification result.

[0149] In step (4), the defect density calculation and the prediction of the optimal cutting path are required, including:

[0150] Defect density calculation:

[0151] Divide the crystal into several macroscopic regions (e.g., in units of cutting planes).

[0152] For each region, count the number of defect voxels N of each type and calculate the defect density D:

[0153]

[0154] Where A is the area of ​​the cutting plane of the region.

[0155] Spatial distribution analysis:

[0156] Using the defect density data, a defect spatial distribution map is drawn to show the defect density in different areas.

[0157] Prediction of the best cutting plane:

[0158] Dataset preparation: The historical defect distribution data and cutting effect data are combined into a training set to train the convolutional neural network model.

[0159] Model input: defect spatial distribution data of the current crystal.

[0160] Model output: Predicted optimal cutting plane location and angle to minimize defects and maximize quality after cutting.

[0161] Determination of cutting parameters: Calculate the specific coordinates of the cutting path based on the prediction results of the optimal cutting plane.

[0162] Based on the material properties of the crystal and the performance of the cutting equipment, the cutting parameters such as cutting speed v, laser power P, pulse frequency f, etc. are determined.

[0163] In step (5), performing cutting according to the optimal cutting path includes,

[0164] Cutting path loading:

[0165] The optimal cutting path and cutting parameters are input into the control unit of the cutting system.

[0166] Cutting parameter settings:

[0167] According to the cutting parameters, adjust the laser output power, focus position, moving speed, etc.

[0168] Real-time feedback control: During the cutting process, the sensors of the cutting system (such as optical sensors, temperature sensors) monitor the cutting status in real time.

[0169] If an abnormality is detected (such as cutting deviation, overheating), the cutting parameters are adjusted through the feedback mechanism to ensure cutting quality.

[0170] Cutting execution: The cutting system cuts the silicon carbide crystal according to the set path and parameters to obtain high-quality wafers.

[0171] Defect identification example:

[0172] Microtubule defects:

[0173] Features: Appears as elongated cavities or low-density areas on CT images.

[0174] Identification: The CNN model identifies these features and outputs a higher probability of microtubule defects.

[0175] Carbon Wrap Defects:

[0176] Characteristics: Appears as small particles or masses of high density in CT images.

[0177] Identification: The CNN model identifies these high-density areas and outputs a higher probability of carbon encapsulation defects

[0178]

[0179] Polycrystalline area:

[0180] Features: Irregular texture and grain boundaries appear in CT images.

[0181] Identification: The CNN model outputs a higher probability of defects in polycrystalline areas through texture analysis.

[0182] Defect density calculation example:

[0183] For a certain area, count the number of microtubule defect voxels N 微管 =500, area A = 1cm 2 , then the micropipe defect density is:

[0184]

[0185] Example of optimal cutting plane prediction:

[0186] Input: Defect spatial distribution data of the current crystal.

[0187] Model prediction: The CNN model calculates the total defect density of different cutting planes and selects the plane with the smallest total defect density as the optimal cutting plane.

[0188] Output: The position of the optimal cutting plane z = z 0 and angle θ = θ 0 .

[0189] Example of determining cutting parameters: According to the crystal material characteristics and cutting equipment performance, set the cutting speed v = 2 mm / s, laser power P = 1.5 GW, and pulse frequency f = 500 kHz.

[0190] Cutting execution example: The cutting system starts cutting according to the set path and parameters.

[0191] Real-time monitoring of the temperature T and stress σ in the cutting area. If T exceeds the set value T max Or σ exceeds the set value σ max , then reduce the laser power or slow down the cutting speed.

[0192] The present invention achieves the following effects:

[0193] Defect identification: X-ray CT scanning combined with CNN models can accurately identify different types of defects, which has been widely used in the fields of material science and medical imaging.

[0194] Defect density calculation: It is a commonly used method to calculate the defect density by counting the classification results of voxels.

[0195] Optimal cutting plane prediction: Using machine learning models to predict the optimal cutting plane based on the spatial distribution of defects can effectively improve chip quality.

[0196] Cutting parameter determination and execution: Combining material characteristics and equipment performance, setting cutting parameters, and ensuring cutting quality through real-time feedback control are key technologies in modern precision machining.

[0197] Beneficial effects of the method steps:

[0198] Step (1) cleans the surface of the silicon carbide crystal by using ethanol or isopropanol to remove contaminants, and then blows it dry with nitrogen to prevent oxidation, thereby ensuring the cleanliness of the crystal surface, preventing the influence of oxidation on subsequent scanning and cutting, and improving data accuracy and cutting quality.

[0199] Step (2) fixes the silicon carbide crystal on a crystal fixing device with the function of rotating and moving up and down, and sets the rotation speed to 1 revolution per minute, thereby realizing multi-angle and multi-directional scanning. By acquiring X-ray data at different angles, the data source for three-dimensional reconstruction is enriched, and the accuracy of the reconstructed model is improved.

[0200] Step (3) uses the Feldkamp-Davis-Kress (FDK) cone beam CT reconstruction algorithm to process the X-ray data and reconstruct the 3D model of the silicon carbide crystal, which accurately reflects the structure and defect information inside the crystal. The deep learning neural network trained with labeled defect data is used to accurately identify and classify defect types such as microtubes, carbon packaging, and polycrystalline areas, and calculate the defect density D of each area. By grading the defect density, an intuitive defect distribution map is provided, which provides a basis for optimizing the cutting path.

[0201] Step (4) uses a machine learning algorithm to analyze the spatial distribution of defects, predict the best cutting plane, and calculate the optimal cutting path by minimizing the total defect density in the resulting wafer. Based on the defect data and material properties, the cutting parameters of each cutting path segment are determined to ensure the targeted and efficient cutting process.

[0202] Step (5) sets the cutting system to perform cutting according to the optimal cutting path, and ensures the stability and efficiency of the cutting process through real-time monitoring and adjustment. Through the optimized cutting path and parameter adjustment, the defects in the wafer are minimized, the quality of the wafer is improved, and the market thickness and quality requirements are met or exceeded.

[0203] Step (6) collects the cut silicon carbide wafers. Due to the optimized cutting path and real-time parameter adjustment, the wafer defects are minimized to reach the level of Class A defects. If necessary, secondary 3D defect detection and secondary cutting can be performed to further improve the wafer quality and ensure the high quality of the product.

[0204] In summary, the present invention achieves efficient and high-precision cutting of silicon carbide crystals by combining 3D defect detection, artificial intelligence algorithms, machine learning, and real-time adjusted laser cutting technology. Each step is closely connected, and the output rate of low-defect wafers is maximized through precise defect detection, intelligent cutting path optimization, and real-time cutting parameter adjustment. The system and method significantly improve wafer quality, reduce production costs, and improve production efficiency, and have important application value and broad market prospects.

[0205] In order to verify the effectiveness and advantages of the proposed silicon carbide crystal cutting system and method based on 3D defect detection, the following specific experiments were designed and conducted.

[0206] Experimental preparation and detailed implementation process:

[0207] A high-purity silicon carbide (SiC) single crystal with a diameter of 200 mm and a thickness of 20 mm was selected as the experimental sample. First, the surface of the SiC crystal was cleaned with anhydrous ethanol and wiped thoroughly to remove organic matter and particulate contaminants on the surface. Then, high-purity nitrogen was used to purge the crystal surface to ensure thorough drying to prevent oxidation and water vapor from affecting subsequent testing.

[0208] The processed SiC crystal is fixed on a crystal fixing device with high-precision rotation and up and down movement functions. The rotation speed of the device is set to 1 revolution per minute (1rpm) to ensure that comprehensive internal information of the crystal can be obtained from different angles during the scanning process. The micro-focus high-energy X-ray tomography CT scanning device is started, and the device parameters are set to an acceleration voltage of 250kV, a focal spot size of less than 10 microns, an exposure time of 1000ms, and a spatial resolution of 80 microns. While the crystal rotates, X-rays irradiate the crystal from different angles to obtain complete X-ray data signals.

[0209] The collected X-ray data signals are input into the 3D data processing module. First, the Feldkamp-Davis-Kress (FDK) cone-beam CT reconstruction algorithm is used to reconstruct the 2D cross-sectional image sequence into 3D to form the 3D volume data of SiC crystal. In order to improve the data quality, the Gaussian filtering method is used to denoise the reconstructed 3D data to eliminate the random noise and artifacts generated during the scanning process.

[0210] Next, the 3D volume data is divided into voxels, each of which represents a tiny cubic area in the crystal. For each voxel, parameters such as grayscale value and texture features are extracted, which reflect the subtle differences in the internal structure of the crystal. The extracted features are input into a pre-trained convolutional neural network (CNN) model, which is trained based on a large amount of labeled defect data and has high-accuracy defect classification capabilities.

[0211] The CNN model outputs the probability value P of each voxel belonging to a specific defect type. defect (x, y, z), where x, y, z are the positions of the voxels in three-dimensional space, and defect types include microtubes, carbon inclusions, polycrystalline regions, etc. For each voxel, the defect type with the highest probability is selected as its classification result.

[0212] The crystal is divided into several macroscopic regions according to the thickness of 0.5 mm. The defect density D of different regions is calculated for each region using the formula D=N / A, where N is the product of the number of defects detected in area A. Defects are divided into four levels according to the defect density per unit area:

[0213] Grade A: D<1000 pieces / cm 2 ;

[0214] Grade B: D is 1000-3000 pieces / cm 2 between;

[0215] C grade: D is 3000-8000 pieces / cm 2 between;

[0216] D grade: D>8000 pieces / cm 2 .

[0217] Using the above results, a 3D defect image is generated, and different defect types and levels are represented by different colors, which facilitates intuitive observation of the spatial distribution of defects inside the crystal.

[0218] The 3D defect image data is input into the cutting data processing module. The CNN algorithm trained with historical defects and cutting data is used to analyze the spatial distribution of defects and predict the optimal cutting plane. The optimal cutting path is calculated by minimizing the total defect density in the resulting wafer. The objective function is:

[0219]

[0220] Among them, D 总 is the total defect density, Di is the defect density of the i-th layer, and hi is the thickness of the i-th layer. Based on the defect data and material properties, the cutting parameters for each cutting path segment are determined, including cutting depth, speed, and laser power settings.

[0221] Set up the cutting system as a laser cutting system that can adjust parameters in real time. Input the cutting parameters from the cutting data processing module to the cutting system, and initially set them to a cutting line width of 0.15mm, a cutting speed of 3mm / s, and a laser peak power of 1.5GW. Start the cutting process, and the laser cuts according to the predetermined optimal cutting path. During the cutting process, in-situ sensors (such as optical sensors and temperature sensors) monitor the cutting status in real time. If an abnormality is detected (such as cutting deviation, overheating, etc.), the cutting parameters such as laser power, focus position and speed are adjusted through a feedback mechanism to ensure cutting quality.

[0222] Finally, the cut silicon carbide wafers are collected. Due to the optimized cutting path and real-time parameter adjustment, the defects of the wafers are minimized while meeting or exceeding market quality requirements.

[0223] Table 1: Comparison of performance of silicon carbide wafers cut by conventional method and method of the present invention

[0224]

[0225]

[0226] It can be seen from the data in Table 1 that the wafers cut by the method of the present invention (wafer numbers 4-6) are significantly better than the wafers cut by the traditional method (wafer numbers 1-3) in various performance indicators.

[0227] First, in terms of average total dislocations, the average defect densities of wafers cut by conventional methods were 4560, 3780, and 4210 / cm2, respectively. 2 , indicating that there are many defects inside the wafer. The average defect density of the wafer cut by the method of the present invention is significantly reduced, which is 1530, 1800 and 1720 / cm respectively. 2 This shows that the method of the present invention can effectively detect and avoid defective areas inside the crystal, significantly improving the quality of the wafer.

[0228] In terms of wafer qualification rate, the yield rate of the conventional method is between 80% and 85%, while the yield rate of the method of the present invention is increased to more than 95%. This means that the method of the present invention not only greatly improves the quality of wafers, but also significantly reduces the scrap rate in the production process, improves production efficiency, and reduces production costs.

[0229] Based on the above analysis, the method of the present invention achieves accurate identification and positioning of internal defects of silicon carbide crystals by introducing 3D defect detection and artificial intelligence optimization technology. Using machine learning algorithms such as CNN, the optimal cutting path can be predicted to avoid defect-intensive areas as much as possible. During the cutting process, the cutting parameters are adjusted through real-time monitoring and feedback to ensure the stability and consistency of the cutting quality.

[0230] In contrast, traditional cutting methods lack effective detection of internal crystal defects and cutting path optimization, resulting in high wafer defect density and low wafer pass rate, limiting the application and market competitiveness of silicon carbide wafers.

[0231] Therefore, the data of this embodiment fully demonstrates the significant advantages of the method of the present invention in improving wafer quality, optimizing cutting efficiency, reducing production costs, etc., embodies the innovativeness and practical value of the present invention, and plays an important role in promoting the production and application of silicon carbide wafers.

[0232] It will be appreciated that in the development of any actual implementation, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but will be a routine task of design, fabrication, and production for those of ordinary skill having the benefit of this disclosure without undue experimentation.

[0233] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A silicon carbide crystal cutting system based on 3D defect detection, characterized in that: include, X-ray tomography (CT) scanning equipment, used to perform all-round, high-resolution 3D scanning of silicon carbide crystals to obtain the structure and defect information inside the crystals; Silicon carbide crystal fixing equipment, used to fix silicon carbide crystals, with the functions of rotation and vertical movement up and down, to ensure that the position of the crystal is accurately controlled during scanning and cutting; The 3D data processing module is connected to the X-ray tomography CT scanning device, and includes a defect data module, a defect data comparison and analysis module, and an image processing and generation module. After the 3D data processing module obtains the X-ray data signal, it obtains a 3D defect image of the silicon carbide crystal. The cutting data processing module is connected to the 3D data module. After receiving the 3D defect data and images, the cutting data module processes and calculates to form the best cutting path method; The cutting system is connected to the cutting data processing module, sets the cutting process parameters according to the optimal cutting path, and finally forms the silicon carbide wafer.

2. The silicon carbide crystal cutting system based on 3D defect detection according to claim 1, characterized in that: The X-ray tomography CT scanning equipment is a micro-focus high-energy X-ray tube. The acceleration voltage range of the X-ray light source is 200-350kV, the focal size of the light source is ≤10 microns, the exposure time is ≥1000ms, and the spatial resolution is ≤80 microns.

3. The silicon carbide crystal cutting system based on 3D defect detection according to claim 1 or 2, characterized in that: The defect data module includes the micro-pipe defect data module, the carbon package defect data module, and the polycrystalline / polytype defect data module; the defect data comparison and analysis module classifies different defect data; the image generation module assigns different defect types different colors based on the comparison and analysis results, and the data within the normal range is represented by black.

4. The silicon carbide crystal cutting system based on 3D defect detection according to claim 3, characterized in that: Based on 3D defect image data, the cutting data processing module contains a defect distribution analysis module and a cutting path analysis module. The cutting path analysis module of the cutting data processing module is provided with 4 silicon carbide defect level divisions, and the Class A defect index is <1000 / cm 2 、Class B level defect index 1000-3000 / cm 2 ; Class C defect index 3000-8000 / cm 2 ; Class D defect index >8000 / cm 2 .

5. A cutting method of a silicon carbide crystal cutting system based on 3D defect detection according to any one of claims 1 to 4, characterized in that: Step 1: Clean the surface of the silicon carbide crystal; Step 2: Place the silicon carbide crystal on a fixed device, set the rotation or movement speed, and the X-ray tomography CT scan emits X-rays, and the silicon carbide crystal rotates or moves up and down to obtain X-ray data signals; Step 3: The X-ray data signal is input into the 3D data processing module, and the X-ray data is processed and analyzed according to the difference in X-ray absorption or scattering of different defect parts, and finally a 3D defect image is formed through the image generation module; Step 4: The 3D defect image data is input into the cutting data module, which performs data analysis based on the four silicon carbide defect levels to obtain the best cutting path; Step 5: The cutting system performs cutting according to the optimal cutting path and selects appropriate cutting parameters to ultimately obtain silicon carbide wafers of different specifications.

6. The cutting method of the silicon carbide crystal cutting system based on 3D defect detection according to claim 5, characterized in that: In step 4, the cutting data module performs data analysis based on the four silicon carbide defect levels, and sets the cutting path to pass through the defect-dense area to the greatest extent, and the first cutting ultimately obtains silicon carbide wafers of different thicknesses with minimized surface defects; For silicon carbide wafers of different thicknesses with minimized surface defects obtained by the first cutting, a cutting path analysis can be further performed, and a second cutting can be performed to obtain silicon carbide wafers with minimized defects and meeting thickness requirements.

7. The cutting method of the silicon carbide crystal cutting system based on 3D defect detection according to claim 6, characterized in that: In step 4, the cutting data module performs data analysis based on the four silicon carbide defect levels, and sets the thickness of the final cut silicon carbide wafer to be consistent, which is 0.35 mm or 0.5 mm, and finally obtains the optimal cutting path with consistent wafer thickness.

8. The cutting method of the silicon carbide crystal cutting system based on 3D defect detection according to claim 7, characterized in that: The X-ray data signal is input into the 3D data processing module, which specifically includes: The X-ray data is processed using the FDK cone-beam CT reconstruction algorithm to reconstruct a 3D model of the silicon carbide crystal, where the reconstructed voxel intensity I(x, y, z) represents the attenuation coefficient of each point in the crystal; Artificial intelligence algorithms such as deep learning neural networks trained with labeled defect data are used to analyze the 3D model and identify defect types such as microtubes, carbon packaging, and polycrystalline areas. The neural network outputs the defect classification probability P for each voxel. defect (x,y,z).

9. The cutting method of the silicon carbide crystal cutting system based on 3D defect detection according to claim 8, characterized in that: 3D defect image data is input into the cutting data processing module, including: Use machine learning algorithms trained with historical defect and cutting data to analyze the spatial distribution of defects and predict the best cutting plane; The optimal cutting path is calculated by minimizing the total defect density in the resulting wafer, using the formula: Minimize D 总 =∑ i (D i ×h i ) Among them, D 总 is the total defect density, D i is the defect density of the i-th layer, h i is the thickness of the i-th layer; Cutting parameters for each cutting path segment, including cutting depth, speed, and power settings, are determined based on defect data and material properties.

10. The cutting method of the silicon carbide crystal cutting system based on 3D defect detection according to claim 8 or 9, characterized in that: Cutting is performed according to the optimal cutting path, including: Inputting cutting parameters from the cutting data processing module into the cutting system; Start the cutting process, the laser cuts according to the predetermined path, and the initial settings are a cutting line width of 0.15mm, a cutting speed of 3mm / s, and a laser peak power of 1.5GW; The cutting process is continuously monitored, and cutting parameters such as laser power, focus and speed are adjusted in real time based on the feedback from the 3D data processing module and the in-situ sensor to cope with the changes in defect distribution encountered during the cutting process.

Images