A method and apparatus for identifying and processing a specific region of an ingot

Abstract

The present invention relates to the technical field of semiconductor ingot processing, and particularly to a method and device for identifying and processing specific regions of an ingot. The present invention provides a method for identifying and processing specific regions of an ingot. This method uses a light source to irradiate and a camera to collect color pattern images, and combines a region recognition algorithm to automatically identify specific regions in the ingot where cracks have not formed, and perform secondary processing on these regions until the specific regions disappear or meet the processing standards. This method can improve the automation and precision of processing, reduce manual intervention, ensure the uniformity and integrity of ingot cracks, and improve the peeling quality and production efficiency of wafers. At the same time, this method effectively solves the technical problems of identifying and processing specific regions in the prior art and has high industrial application value.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor ingot processing, and particularly to a method and a device for identifying and processing specific regions of an ingot. Background Art

[0002] With the wide application of materials such as silicon carbide in the fields of photovoltaic, semiconductor, etc., the manufacturing and processing of ingots have become one of the key links. During the production process of an ingot, laser processing is usually required to form a crack layer inside the ingot, so as to achieve efficient wafer cutting during the subsequent peeling process. However, during the laser processing process, due to the physical properties of the ingot material itself, uneven doping concentration, and differences in processing process parameters, cracks are not effectively formed in some regions or the crack connections are incomplete. These regions are called "specific regions". If not processed, problems such as wafer cracking and edge chipping are likely to occur during the subsequent wafer peeling process, affecting product quality and production efficiency.

[0003] In the prior art, ingot processing usually relies on a combination of automatic control of the laser system and manual operation. The observation and judgment of the color pattern image by humans are subjective and uncertain, making it difficult to achieve high-precision and high-efficiency identification and processing. In addition, when the traditional light source and camera combination are used to identify the cracks in the ingot, limited by the equipment accuracy and light conditions, the regions without cracks cannot be accurately identified, resulting in incomplete processing of specific regions. To solve these problems, a more precise and intelligent method for identifying and processing specific regions of an ingot is needed. Summary of the Invention

[0004] To solve the above technical problems, the purpose of the present invention is to provide a method for identifying and processing specific regions of an ingot. This method uses a light source to irradiate and a camera to collect color pattern images, and combines a specific region identification algorithm to automatically identify the specific regions in the ingot where cracks are not formed, and perform secondary processing on these regions until the specific regions disappear or meet the processing standards. This method can improve the automation and precision of processing, reduce manual intervention, ensure the uniformity and integrity of the ingot cracks, and improve the peeling quality and production efficiency of the wafers. At the same time, this method effectively solves the technical problems of identifying and processing specific regions in the prior art and has high industrial application value.

[0005] To achieve the above purpose, the present invention adopts the following technical solutions:

[0006] A method for identifying and processing specific regions of an ingot, comprising the following steps:

[0007] 1) After the front-end ingot laser processing forms a peeling zone, use a light source to irradiate the surface of the ingot and use a camera to collect color pattern images;

[0008] 2) Preprocess the collected color pattern images;

[0009] 3) Identify specific regions in the ingot through a specific region recognition algorithm, where the specific regions are those without observed color patterns after the previous laser processing of the ingot;

[0010] 4) Perform secondary processing on the specific regions until these specific regions disappear or meet the processing standards.

[0011] Preferably, the light source includes a surface light source or a strip light source. The camera needs to be placed on the optical path after the light irradiated by the light source is reflected by the ingot crack, and this spatial relationship needs to be fine-tuned according to the camera frame size to ensure clear capture of the color patterns.

[0012] Preferably, the specific regions are those without showing color patterns after the previous laser processing due to the non-formation of cracks or unconnected cracks, and these regions are likely to appear in the facet regions or low-resistivity regions of the ingot.

[0013] Preferably, when using a strip light source, image stitching includes the following steps:

[0014] a) First, determine the stitching starting point by collecting the center point coordinates of the first image;

[0015] b) Drive the ingot or the strip light source and the camera to move through a motor, and the moving distance is less than or equal to the predetermined distance d;

[0016] c) Crop and stitch multiple images through coordinates to finally form a complete color pattern image.

[0017] More preferably, during the stitching process, image fusion processing is performed for each stitching point position to eliminate stitching traces and ensure the continuity and integrity of the color pattern information.

[0018] Preferably, the specific region recognition includes the following steps:

[0019] a) Segment the preprocessed color pattern images through a deep learning image segmentation algorithm;

[0020] b) The image segmentation model is trained using a fully convolutional neural network (FCN). Use the labeled image dataset for model training, and optimize the model by adjusting hyperparameters such as the learning rate and batch size.

[0021] Preferably, in the model training stage, the OpenCV library is used for image preprocessing, and the PyTorch framework is used for model training. After training, the unlabeled color pattern images are input into the model to obtain the segmented specific region results.

[0022] Preferably, in the specific area processing step, the following parameters are adjusted according to the process requirements: the light intensity of the light source; the depth of laser processing; the speed and spacing of laser scribing; the path of laser scribing; by adjusting the above parameters, the specific area is processed to ensure the generation and connection of cracks, avoiding the risk of splitting or edge collapse during subsequent peeling.

[0023] Preferably, after the processing of the specific area is completed, the color pattern image of the ingot is collected and identified again to determine whether the specific area has completely disappeared. If not, the identification-processing cycle steps are repeated until the processing is completed.

[0024] Furthermore, the present invention also discloses a device for identifying and processing a specific region of an ingot for implementing the method, the device comprising:

[0025] The detection module composed of a light source and a camera collects the image of the color pattern of the crystal ingot;

[0026] A specific area recognition module based on deep learning, used to segment and identify specific areas;

[0027] Specific area processing module, adjust laser processing according to process parameters;

[0028] The result display module returns the recognition and processing results to the client through network communication.

[0029] The present invention adopts the above-mentioned technical solution, mainly through the steps of light source irradiation, camera acquisition of color pattern images, image preprocessing, specific area recognition algorithm and secondary processing, to solve the technical problem that specific areas cannot be effectively identified and processed during ingot processing in the prior art. The technical effects of the present invention are reflected in the following aspects:

[0030] 1. Improved the accuracy of specific area recognition: By using light source illumination and camera to collect color pattern images, combined with a specific spatial relationship configuration, the color pattern information on the surface of the ingot can be clearly captured, especially the characteristic image of crack reflection light. Through image preprocessing technology (such as image stitching, cropping, etc.) and specific area recognition algorithms, specific areas where no cracks are formed or cracks are not connected after laser processing can be accurately identified, thereby greatly improving the recognition accuracy.

[0031] 2. Realize intelligent secondary processing of specific areas: After identifying a specific area, the present invention uses secondary processing to adjust laser process parameters (such as light intensity, depth, scribing path, etc.) to perform targeted processing on the specific area to ensure that a complete crack is formed in the area until the area disappears or meets the preset processing standards. This intelligent secondary processing process effectively reduces the incidence of problems such as cracking and edge collapse, and ensures the processing quality of the ingot.

[0032] 3. Reduced manual intervention and improved production efficiency: Traditional identification and repair of specific areas on ingots often rely on manual observation and manual correction. However, through the automated processes of image acquisition, identification, and secondary processing in the present invention, the reliance on manual operations is significantly reduced, the errors caused by subjective judgment are decreased, the stability and consistency of the processing process are improved, and thus the production efficiency is remarkably enhanced.

[0033] 4. Optimized wafer peeling quality: Through the method of the present invention, the uniformity and integrity of internal cracks in the ingot can be ensured, and defects in wafer peeling caused by unformed or unconnected cracks, such as chip cracking and edge chipping, can be avoided. After the specific area undergoes secondary processing, it can better meet the requirements of subsequent wafer peeling, thereby improving the overall quality of the wafer.

[0034] 5. Strong adaptability: The present invention is not only applicable to conventional ingot materials but can also flexibly adapt to the processing requirements of different ingots by adjusting laser processing parameters and identification algorithms according to the characteristics of different materials (such as conductive SiC, etc.), thus having broad industrial application prospects.

[0035] In summary, through the automated and intelligent method of specific area identification and processing, the present invention improves the processing quality of ingot cracks, reduces manual intervention, significantly enhances the production efficiency, and ensures the quality and stability during the wafer peeling process, having important technical advantages and application values. Description of the Drawings

[0036] Figure 1 It is an example of the included angle between the cleavage plane of conductive SiC and the crystal surface.

[0037] Figure 2 It is a schematic diagram of the crack generation process of conductive SiC.

[0038] Figure 3 It is a schematic diagram of the light reflected during the crack generation process of conductive SiC.

[0039] Figure 4 It is Example Ⅰ of the spatial relationship among the light source, camera, and ingot.

[0040] Figure 5 It is Example Ⅱ of the spatial relationship among the light source, camera, and ingot.

[0041] Figure 6 It is the system flow chart of the method of the present invention.

[0042] Figure 7 It is the schematic diagram of image stitching of the bar light source.

[0043] Figure 8 It is the structural block diagram of a device for identifying and processing characteristic areas of an ingot according to the present invention. Detailed implementation mode

[0044] Combined with the embodiments of the present invention below, the technical solutions in the embodiments will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present invention.

[0045] The present invention is based on a method for identifying and processing a specific area of a crystal ingot after forming a peeling band by laser processing in the previous process. As Figure 1 shown, after forming a peeling band by laser processing on the previous crystal ingot, there is an angle (for example, 4°) between the cleavage plane of conductive SiC and the crystal surface, and the intersection line of the two is the crystal orientation. As Figure 2 shown, during laser processing, laser-induced modified particles are generated inside the SiC crystal ingot, and cracks are generated along the cleavage plane direction. The cracks will not extend infinitely along the cleavage plane, but will connect with the previous and the next cracks after extending a certain length. As Figure 3 shown, after the cracks are connected, a surface light source, a line light source or other light sources are used to irradiate the surface of the crystal ingot, and the light reflected by the crack surface will be reflected with the perpendicular line of the cleavage plane as the normal.

[0046] As Figure 4 、 5 shown is an example of the spatial relationship among the light source, the camera and the crystal ingot. After forming a crack layer inside the crystal ingot by laser, in order for the camera to capture the light reflected by the crack after the light source irradiates, the following conditions need to be met: The camera needs to be placed along the propagation direction of the reflected light after the light irradiated by the light source is reflected by the crack of the crystal ingot, and fine-tuning is carried out according to the frame size of the camera. Because the actual lengths of the cracks generated inside the crystal ingot after laser processing are different, the intensities of the light reflected by the cracks are different, and the result observed by the camera or the naked eye is colored stripes with different brightnesses.

[0047] Therefore, the specific area defined in this application is defined as follows: After the previous crystal ingot is processed by laser, affected by factors such as inappropriate selection of various process parameters and uneven doping concentration of the crystal ingot itself, some areas have formed cracks and colored stripes are observed, but there are still some areas that may not show colored stripes because there are no cracks or the cracks are not connected. These areas that do not show colored stripes because there are no cracks or the cracks are not connected are defined as specific areas. For conductive SiC, specific areas are likely to appear in the facet area and its vicinity. Because during the production process of conductive SiC, due to different growth modes in the facet area and the non-facet area, physical properties such as resistivity are different, and there are also significant differences in the absorption of laser light. The facet area is prone to having no cracks or the cracks not being connected and thus not showing colored stripes after laser processing.

[0048] As Figure 6As shown, a method for identifying and processing a specific area of an ingot specifically includes the following steps:

[0049] 1. Form a peeling zone through pre-stage laser processing

[0050] First, perform pre-stage laser processing on the surface of the ingot. The laser generates modified particles inside the ingot and forms cracks along the cleavage plane to form a peeling zone. During this process, the cracks do not expand indefinitely but connect with the front and rear cracks after a suitable length to form a complete crack layer, which is used to facilitate the subsequent peeling of the wafers. However, in some areas, due to the influence of material properties or processing parameters, a complete crack may not be formed, and at this time, a "specific area" will be generated.

[0051] 2. Acquisition of the color pattern image

[0052] After the pre-stage laser processing is completed, irradiate the surface of the ingot by setting a suitable light source (such as a surface light source or a strip light source), and use a camera to acquire the color pattern image of the ingot. A certain spatial relationship needs to be formed among the camera, the light source, and the ingot to ensure that after the light source irradiates the ingot, the light reflected by the cracks can be accurately captured by the camera. Especially when the light source is a strip light source, drive the movement of the ingot or the light source and the camera through a motor to gradually acquire the entire color pattern information of the ingot.

[0053] 3. Preprocessing of the color pattern image

[0054] The acquired color pattern image needs to be preprocessed, specifically including operations such as image stitching and cropping. The strip light source can only illuminate part of the area of the ingot, and the color pattern image is composed of multiple local images, and image stitching needs to be performed through a stitching algorithm.

[0055] As Figure 7 shown, when using a strip light source, due to the range of the light source, only part of the area of the ingot surface can be illuminated, such as Figure 7 the rectangular area shown in the first image in a. At this time, the camera can only detect the color patterns in the rectangular area, and the rest of the ingot is darker in color. Therefore, it is necessary to preprocess the photos of the ingot taken under the strip light through image stitching to present the color pattern situation of the entire surface of the ingot, and then use it for the recognition and calculation of subsequent algorithms.

[0056] The image stitching process includes: As Figure 7 shown in a, confirm the starting point of the world coordinate through the center point coordinates of the first image and the stitching point position, and drive the movement of the ingot through a motor (it can also drive the movement of the strip light source and the camera). The condition to ensure that the color pattern information is not missing is that the movement distance each time ≤ d, Figure 7 as shown in b for the example of moving = d. After collecting the images at all positions, calculate the stitching point positions of the acquisition areas of each image, perform cropping and stitching, and then the complete color pattern image of the ingot can be obtained. Figure 7c in

[0057] 4. Identification of Specific Regions

[0058] The pre - processed color pattern image will be analyzed by a specific region recognition algorithm. The specific region refers to the area where cracks have not formed or the cracks are not connected after the previous laser processing, and these areas do not show color pattern features in the color pattern image. To improve the recognition accuracy, the present invention uses an image segmentation algorithm based on deep learning, such as the fully convolutional neural network (FCN), etc., for automatic identification of specific regions. This algorithm first trains the color pattern image with a large amount of historical data, and then inputs the newly acquired image into the trained model to automatically identify the specific region and output the result.

[0059] The goal of specific region recognition is to segment specific regions and non - specific regions. The specific region recognition algorithm of the present invention basically includes four stages: image annotation, training environment construction, model training, and model prediction;

[0060] In the image annotation stage, for the previously saved complete ingot image after splicing, specific regions are manually or automatically marked;

[0061] In the training environment construction stage, necessary software libraries and dependency libraries are installed according to the actual segmentation algorithm used to construct the training environment;

[0062] In the model training stage, the marked images are input into the corresponding image segmentation model for training;

[0063] In the model prediction stage, the color pattern image of the ingot not used for training is input into the trained model to obtain the segmentation result.

[0064] In this embodiment, the present patent uses the fully convolutional neural network as the segmentation model. First, collect multiple complete ingot splicing images, and use manual and automated tools to accurately label the specific regions in the images at the pixel level; according to the selected image segmentation algorithm, build the corresponding training environment, mainly using the PyTorch deep learning framework and the OpenCV image processing library, select the CNN image segmentation model, divide the labeled image dataset into a training set and a validation set, and input them into the model for training. During the training process, the model performance is optimized by adjusting hyperparameters such as the learning rate and batch size.

[0065] The specific steps for predicting the color pattern image of a new ingot using the trained model are as follows:

[0066] 1) Input the ingot image to be predicted;

[0067] 2) Pre - process the image, including scaling, normalization, etc.;

[0068] 3) Input the pre - processed image into the trained model;

[0069] 4) The model outputs the segmentation result, i.e., the position and shape of the specific area;

[0070] 5) Post-process the segmentation result, including boundary smoothing, small area filtering, etc.;

[0071] 6) Overlay the processed result on the original image to visually display the recognized specific area.

[0072] 5. Secondary processing of the specific area

[0073] After identifying the specific area, it is necessary to perform secondary processing on this area to ensure the formation and connection of cracks and avoid wafer breakage or edge chipping during subsequent peeling. The secondary processing is carried out by laser, and the following parameters are adjusted according to the actual process requirements: the light intensity of the light source; the depth of laser processing; the speed of laser scribing; the laser line spacing; the path of laser scribing. After the processing is completed, the color pattern image is collected and recognized again to determine whether the specific area disappears. If the color pattern is complete, it means the processing is completed; if the specific area still does not meet the requirements, the process of recognition and secondary processing is continued until the specific area disappears or meets the processing standard.

[0074] As Figure 8 shown, a device for identifying and processing characteristic areas of an ingot according to the present invention is composed of a specific area detection module, a specific area recognition module, a specific area processing module, and a result display module.

[0075] Specific area detection module: This module is composed of a light source and a camera that form a certain spatial relationship with the ingot, and the camera needs to be placed along the propagation direction of the reflected light after the light from the light source irradiates the ingot crack. When the light source is a large enough surface light source, the entire ingot is illuminated by the surface light source and photographed; when the light source is a strip light source, the ingot is driven to move by a motor (or the strip light source and the camera can also be driven), and each time it moves a distance ≤d until the entire ingot is photographed, trimmed and spliced to obtain a complete color pattern image of the ingot.

[0076] Specific area recognition module: Train and deploy a trained model on the server side, input the photographed color pattern image into the trained model through socket communication, obtain the segmentation result and return it to the client.

[0077] Result display module: After identifying the specific area, transmit the recognition result back to the client through socket communication and display it.

[0078] Specific area processing module: Adjust process parameters including light intensity, processing depth, laser line spacing, speed of laser scribing, path of laser scribing, etc. according to the actual situation, and perform additional processing on the specific area according to the adjusted result of the process parameters.

[0079] Experimental Example

[0080] To prove the effectiveness of the method for identifying and processing specific regions of the ingot in the present invention, a series of experiments were designed to test the effect of the present invention in actual ingot processing and compare it with traditional methods.

[0081] 1. Experimental Object: Conduct experiments using conductive SiC standard ingot materials. Among them, the doping concentration of the ingot material itself is uneven, and it is difficult to form complete cracks through primary laser processing.

[0082] 2. Testing Equipment: Adopt a light source, camera system, and deep learning segmentation algorithm with the functions of the present invention, as well as laser processing equipment.

[0083] 3. Experimental Content: Process specific regions through the color pattern image acquisition, preprocessing, specific region identification, and secondary processing processes of the present invention, and compare with traditional processing methods that rely on manual identification and repair.

[0084] 4. Evaluation Criteria:

[0085] Recognition Accuracy: The percentage of accurately identifying specific regions.

[0086] Processing Time: The time required to complete the identification and secondary processing of specific regions of each ingot.

[0087] Chip Splitting Rate: The probability of chip splitting or edge chipping during the subsequent peeling process of each ingot.

[0088] Product Qualification Rate: The proportion of wafers meeting the standard quality requirements after processing.

[0089] Table 1 Experimental Data

[0090]

[0091] 6. Analysis of Experimental Results

[0092] 1) Recognition Accuracy: The method of the present invention uses a high-precision camera and deep learning segmentation algorithm, and the recognition accuracy of specific regions is as high as 99.5%, which is better than 93.4% of traditional methods that rely on manual observation. Manual observation is limited by lighting conditions and the experience of observers, and it is easy to miss or misjudge some specific regions. The present invention can perform high-precision image processing on the ingot surface, greatly improving the recognition accuracy.

[0093] 2) Processing Time: The average processing time for each ingot using the method of the present invention is 5 minutes, which is 15 minutes shorter than traditional methods. The reason is that the present invention reduces the manual judgment and intervention process through an automated image recognition and processing flow, accelerating the processing speed of specific regions.

[0094] 3) Split rate: The present invention significantly reduces the split rate, which is only 1.0% in the experiment, while the traditional method has a split rate of 7.5%. This is because the present invention can accurately identify and reprocess the area where no cracks have formed, ensuring the uniformity and integrity of the cracks and avoiding the generation of splits and edge collapse during the wafer peeling process.

[0095] 4) Product qualification rate: The qualified rate of wafer processing products of the present invention reaches 99.0%, which is significantly better than the 92.5% of the traditional method. The higher qualified rate is due to the higher recognition accuracy and the reduction of the cracking rate, which ensures the quality of the final processed wafer.

[0096] The above is a description of the embodiments of the present invention. Through the above description of the disclosed embodiments, professionals and technicians in the field can implement or use the present invention. Various modifications to these embodiments will be apparent to professionals and technicians in the field. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown in this article, but will conform to the widest range consistent with the principles and novelties disclosed herein.

Claims

1. A method for identifying and processing a specific region of an ingot, characterized in that, The following steps are involved: 1) After the previous laser processing of the ingot forms a peeling belt, a light source is used to illuminate the surface of the ingot and a camera is used to collect a color pattern image; 2) Preprocessing the collected color pattern images; 3) identifying a specific area in the ingot by a specific area recognition algorithm, wherein the specific area is an area where no color fringes are displayed because no cracks are formed or the cracks are not connected after the previous laser processing; 4) Perform secondary processing on a specific area until the specific area disappears or meets the processing standards; In step 3), the trained model is used to predict the new ingot color pattern image. The specific steps are as follows: a) Input the ingot image to be predicted; b) Preprocess the image, including scaling and normalization; c) Input the preprocessed image into the trained model; d) The model outputs the segmentation result, i.e. the location and shape of a specific area; e) Post-processing the segmentation results, including boundary smoothing and small area filtering; f) Overlay the processed results on the original image to visually display the identified specific areas; In the specific area processing step, the following parameters are adjusted according to the process requirements: the light intensity of the light source; the depth of laser processing; the speed and spacing of laser scribing; the path of laser scribing; by adjusting the above parameters, the specific area is processed to ensure the generation and connection of cracks, avoiding the risk of splitting or edge collapse during subsequent peeling; after the specific area is processed, the color pattern image of the ingot is collected and identified again to determine whether the specific area has completely disappeared. If not, the identification-processing cycle steps are repeated until the processing is completed.

2. The method according to claim 1, wherein: The light source includes a surface light source or a strip light source, and the camera is placed on the light path of the light irradiated by the light source after being reflected by the cracks of the ingot.

3. The method according to claim 1, characterized in that: When using a bar light source, image stitching consists of the following steps: a) First, determine the stitching starting point by collecting the coordinates of the center point of the first image; b) driving the crystal ingot or bar light source and the camera to move by a motor, and the moving distance is less than or equal to a predetermined distance d; c) Multiple images are stitched together through coordinate cropping to finally form a complete color pattern image.

4. The method according to claim 3, characterized in that: During the stitching process, image fusion processing is performed on each stitching point to eliminate stitching marks and ensure the continuity and integrity of the color pattern information.

5. The method according to claim 1, wherein: The specific area identification comprises the following steps: a) Segment the preprocessed color pattern image using a deep learning image segmentation algorithm; b) The image segmentation model is trained using a fully convolutional neural network (FCN), using a labeled image dataset for model training, and the model is optimized by adjusting the hyperparameters of the learning rate and batch size.

6. The method according to claim 5, wherein: In the model training phase, the OpenCV library is used for image preprocessing, and the PyTorch framework is used for model training. After the training is completed, the unlabeled color pattern image is input into the model to obtain the segmentation results of the specific area.

7. An ingot specific area identification and processing device for implementing the method according to any one of claims 1 to 6, characterized in that, The device comprises: The detection module composed of a light source and a camera collects the image of the color pattern of the crystal ingot; A specific area recognition module based on deep learning, used to segment and identify specific areas; Specific area processing module, adjust laser processing according to process parameters; The result display module returns the recognition and processing results to the client through network communication.

Images