A main roller and diamond wire detection method of a silicon wafer cutting device and a mobile detection terminal

Abstract

The application relates to a main roller of a silicon wafer cutting device and a diamond wire detection method and a mobile detection terminal, and the method comprises the following steps: S1, acquiring wire slot images at the front ends of two main rollers; S2, identifying the empty slot numbers of the two main rollers from the wire slot images; S3, calculating the wrong slot number according to the empty slot numbers of the two main rollers; S4, calculating the wrong slot theoretical distance according to the wrong slot number and the slot pitch; S5, moving to acquire the wrong slot end face images of the two main rollers, then identifying feature points from the wrong slot end face images by using an image calibration and processing method, and then calculating the distance between the two feature points as the wrong slot actual distance by using a pixel principle; and S6, comparing the wrong slot actual distance with the wrong slot theoretical distance to judge whether the main roller installation state and the diamond wire slotting mode are accurate. The application is fast and accurate in slotting, can improve the calculation precision and the judgment accuracy, can reduce the occurrence of the inaccurate main roller installation state and the diamond wire slotting mode, and can reduce the wire skipping occurrence rate.

Description

Technical Field

[0001] This invention belongs to the field of silicon wafer production technology, and particularly relates to a method for detecting the main roller and diamond wire of a silicon wafer cutting equipment, as well as a mobile detection terminal. Background Technology

[0002] Currently, the cutting principle of existing silicon wafer slicing equipment on the market is as follows: diamond wires are distributed and supported by the wire groove on the main roller to form a multi-wire mesh, and then the silicon rod is cut in both directions by acceleration and deceleration through the multi-wire diamond wire mesh.

[0003] The misalignment of the grooves between the main rollers of a silicon wafer slicing machine directly affects the quality of the silicon wafers produced. For example, when the misalignment distance exceeds a threshold, it directly affects the wafer thickness. Therefore, it is necessary to ensure the accurate installation of the main rollers and the accurate grooving method of the diamond wire in the silicon wafer slicing machine.

[0004] Currently, the methods for detecting the accuracy of the main roller installation status and the diamond wire grooving method are all manual. Operators first take photos of the empty grooves using a mobile phone or electron microscope, then manually count the grooves in the enlarged images and compare them with the standard process to confirm consistency. Next, mechanical measuring tools are used to measure the distances in the empty grooves and compare them with the standard process. Finally, based on the comparison results, the accuracy of the main roller installation status and the diamond wire grooving method is determined.

[0005] The existing detection method is time-consuming and subject to too many human factors, resulting in inaccurate measurement results. This reduces the accuracy of the judgment, increases the incidence of inaccuracies in the installation status of the main roller and the diamond wire grooving method of the silicon wafer cutting equipment, and increases the occurrence of wire skipping in the silicon wafer cutting equipment. This increases the probability of cutting thick and thin wafers, resulting in low silicon wafer thickness, precision and consistency. Summary of the Invention

[0006] To address the aforementioned problems in the prior art, this invention provides a method for detecting the main roller and diamond wire in a silicon wafer cutting equipment, as well as a mobile detection terminal.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0008] In a first aspect, the present invention provides a method for detecting the main roller and diamond wire of a silicon wafer cutting equipment, comprising:

[0009] S1. Obtain images of the grooves at the front ends of the two main rollers in the silicon wafer cutting equipment;

[0010] S2. Identify the number of empty slots for the two main rollers from the acquired groove image;

[0011] S3. Calculate the number of misaligned grooves based on the number of empty grooves on the two main rollers;

[0012] S4. Calculate the theoretical distance of the misaligned slots based on the number of misaligned slots and the slot spacing;

[0013] S5. Move to acquire images of the misaligned end faces of the two main rollers, then use image calibration and processing methods to identify feature points from the misaligned end face images, and then use the pixel principle to calculate the distance between the two feature points, which is the actual misalignment distance.

[0014] S6. Compare the actual misalignment distance with the theoretical misalignment distance to determine whether the installation status of the main roller and the grooving method of the diamond wire are accurate.

[0015] By using the aforementioned detection methods for the main roller and diamond wire of the silicon wafer cutting equipment, the groove counting can be performed automatically, quickly, and accurately. This improves the calculation accuracy of the number of misaligned grooves, the theoretical distance between misaligned grooves, and the actual distance between misaligned grooves. It also improves the accuracy of judging the installation status of the main roller and the groove layout of the diamond wire, thereby reducing the occurrence of inaccuracies in the installation status of the main roller and the groove layout of the diamond wire. This effectively reduces the occurrence rate of skipped wires in the silicon wafer cutting equipment and improves the accuracy and consistency of the cut silicon wafer thickness.

[0016] Furthermore, in step S3, the number of misaligned slots is BA;

[0017] In step S4, the theoretical distance of the misalignment is calculated using the following formula:

[0018] (BA)×m;

[0019] Where A and B are the number of empty slots on the two main rollers, respectively, and m is the slot spacing.

[0020] Further, step S2 includes:

[0021] The method of edge detection, edge probing, contour drawing, polygon mapping, and region segmentation is used to comprehensively process the groove image;

[0022] Identification of calibrated objects through feature values;

[0023] The number of slots with diamond wires and slots without diamond wires in the slot image is calculated by data processing, thus obtaining the number of empty slots for the two main rollers.

[0024] Further, in step S5, a mobile detection terminal is used to move and acquire images of the misaligned end faces of the two main rollers. The mobile detection terminal includes a photography device with a macro lens, a guide slide, and a long plate. The long plate is mounted on the pins on the outermost sides of the two main rollers through baffles at both ends. The guide slide is slidably mounted on the long plate. The photography device is mounted on the guide slide and moves with the sliding guide slide to capture images of the misaligned end faces of the two main rollers.

[0025] Secondly, the present invention also provides a mobile detection terminal for the main roller and diamond wire detection method of the above-mentioned silicon wafer cutting equipment. The mobile detection terminal is moved to acquire images of the misaligned end faces of the two main rollers. It includes a camera with a macro lens, a guide slide, and a long plate. The two ends of the long plate have baffles for mounting on the outermost side posts of the two main rollers of the silicon wafer cutting equipment. The guide slide is slidably mounted on the long plate. The camera is mounted on the guide slide and moves with the sliding guide slide to capture images of the misaligned end faces of the two main rollers.

[0026] Furthermore, the long plate is provided with a guide rail along its length, the guide rail is perpendicular to the axis of the two main rollers, and the guide slide is slidably assembled and connected to the guide rail; the macro lens of the photographic equipment mounted on the guide slide faces downward and is located above the misaligned end of the two main rollers.

[0027] The mobile detection terminal described in this invention can automatically and quickly move to collect images of the misaligned end faces of the two main rollers, which is beneficial for the subsequent calculation of the actual misalignment distance, thereby achieving the technical effect of the first aspect mentioned above. Attached Figure Description

[0028] Figure 1 This is a schematic flowchart of the main roller and diamond wire detection method of the silicon wafer cutting equipment described in this invention;

[0029] Figure 2 This is a schematic diagram of the mobile detection terminal in the main roller and diamond wire detection method of the silicon wafer cutting equipment of the present invention;

[0030] Figure 3 This is a schematic diagram of the moving detection terminal acquiring images of the misaligned end faces of the two main rollers in the silicon wafer cutting equipment and diamond wire detection method of the present invention.

[0031] Figure 4 These are images of the grooves on the left and right main rollers in the main roller and diamond wire detection method of the silicon wafer cutting equipment described in this invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0033] like Figure 1 As shown, the present invention provides a method for detecting the main roller and diamond wire of a silicon wafer cutting equipment, comprising the following steps:

[0034] Step S1. Obtain the image of the groove at the front end of the two main rollers in the silicon wafer cutting equipment. Specifically, this is done by taking a picture of the groove at the front end of the two main rollers in the silicon wafer cutting equipment. Alternatively, the image of the groove at the front end of the two main rollers in the silicon wafer cutting equipment can be taken by the mobile detection terminal in step S5.

[0035] Step S2. Identify the number of empty slots on the two main rollers from the acquired groove image; specifically including:

[0036] S2-1. The groove image is comprehensively processed using edge detection, edge probing, contour drawing, polygon mapping, and region segmentation methods. Specifically, the groove image can be analyzed using a trained neural network (e.g., using a 169-layer DenseNet training model) to complete edge detection, edge probing, contour drawing, polygon mapping, and region segmentation.

[0037] S2-2. Identify the calibration object through feature values; where the feature value is the groove width value specified in the process requirements, and the calibration object is a wired groove; specifically, the above neural network identifies the wired groove in the groove image based on the groove width value;

[0038] S2-3. Calculate the number of slots with diamond wires and slots without diamond wires in the slot image through data calculation, and thus obtain the number of empty slots for the two main rollers.

[0039] Step S3. Calculate the number of misaligned slots based on the number of empty slots on the two main rollers; specifically, subtract the number of empty slots on the two main rollers to obtain the number of misaligned slots, i.e., the number of misaligned slots is BA, where A and B are the number of empty slots on the two main rollers, respectively.

[0040] Step S4. Based on the obtained number of misaligned slots and slot spacing, calculate the theoretical distance of the misaligned slots; specifically, the theoretical distance of the misaligned slots is calculated using the following formula.

[0041] (BA)×m;

[0042] In the formula, A and B are the number of empty grooves of the two main rollers, respectively, and m is the groove spacing, which is the value specified in the process requirements.

[0043] Step S5. Move and acquire images of the misaligned end faces of the two main rollers. Then, using image calibration and processing methods, identify feature points from the misaligned end face images (specifically: automatically identify the image after taking a picture, then use a deep learning algorithm to identify the number of grooves, and obtain the actual distance of the misalignment through geometric calculations; the neural network is trained with a large amount of image data, and the trained neural network analyzes the target object, which has accurate feature segmentation and target recognition capabilities compared to traditional image processing methods; at the same time, it supports retraining of the neural network model: according to the complexity of industrial scenarios and detection targets, it supports continued training of the model, thereby forming an industry-specific processing model. That is, a deep convolutional neural network is used to solve the problem of empty grooves and misaligned grooves; the training model uses a 169-layer DenseNet, which accurately identifies empty grooves and misaligned grooves through 7×7 convolutional layers and 3×3 max pooling layers). Then, the distance between two feature points is calculated using the pixel principle (specifically, the distance between two feature points is calculated using CCD imaging measurement technology), which is the actual distance of the misalignment.

[0044] The actual number of misaligned slots is obtained by measuring the absolute difference in the positive X-axis direction of the diamond wire on the left and right main rollers with an instrument and dividing it by the slot spacing.

[0045] In one possible implementation, this step can involve using a mobile detection terminal to acquire images of the misaligned end faces of the two main rollers, such as... Figure 2 and Figure 3 As shown, the mobile detection terminal includes a camera device 100 with a macro lens (such as a smartphone), a guide slide 200, and a long plate 300. The long plate 300 has baffles (not shown) at both ends for mounting on the outermost side posts of the two main rollers (i.e., the left main roller 400 and the right main roller 500) of a silicon wafer cutting device. The guide slide 200 is slidably mounted on the long plate 300. At this time, the guide slide 200 and its associated long plate 300 are parallel to the plane formed by the left and right main rollers. The camera device 100 is mounted on the guide slide 200. Specifically, the long plate 300 has a guide rail 301 along its length. The guide rail 301 is perpendicular to the axial direction of the two main rollers. The guide slide 200 is slidably connected to the guide rail 301, and the macro lens of the camera device 100 mounted on the guide slide 200 faces downwards and is located above the misaligned ends of the two main rollers. The photographic device 100 moves with the sliding guide slide 200 to capture images of the misaligned end faces of the two main rollers. During this process, the guide slide 200 slides from the right main roller of the silicon wafer cutting device to the left main roller.

[0046] Step S6. Compare the actual misalignment distance with the theoretical misalignment distance to determine whether the installation status of the main rollers and the groove layout of the diamond wire are accurate. When the actual misalignment distance equals the theoretical misalignment distance (i.e., the actual number of misalignments equals the relative number of misalignments, where the actual number of misalignments equals the actual misalignment distance divided by the groove spacing), the installation status of the left and right main rollers and the groove layout of the diamond wire are determined to be accurate and without abnormalities. Conversely, when the actual misalignment distance is greater than or less than the theoretical misalignment distance (i.e., the actual number of misalignments is greater than or less than the relative number of misalignments), the installation status of the left and right main rollers and the groove layout of the diamond wire are determined to be inaccurate and abnormal.

[0047] Therefore, the main roller and diamond wire detection method of the silicon wafer cutting equipment described in this invention can automatically, quickly, and accurately count grooves, thereby improving the calculation accuracy of the number of misaligned grooves, the theoretical distance of misaligned grooves, and the actual distance of misaligned grooves. This improves the accuracy of judging whether the installation status of the main roller and the groove layout of the diamond wire are accurate, reduces the occurrence of inaccuracies in the installation status of the main roller and the groove layout of the diamond wire, effectively reduces the skipped wire occurrence rate of the silicon wafer cutting equipment, and improves the accuracy and consistency of the cut silicon wafer thickness.

[0048] The following examples further illustrate the main roller and diamond wire detection method of the silicon wafer cutting equipment described in this invention.

[0049] Example

[0050] The method for detecting the main roller and diamond wire of the silicon wafer cutting equipment described in this embodiment specifically includes the following process:

[0051] 1. Capture images of the grooves at the front ends of the left and right main rollers on the silicon wafer cutting equipment using a mobile phone or other photography device;

[0052] 2. Identify the number of empty slots for the left and right main rollers from the acquired trough images, such as... Figure 4 As shown, the number of empty slots on the edge of the left main roller is B (B=3), and the number of empty slots on the edge of the right main roller is A (A=1).

[0053] 3. Subtract the number of empty slots of the right main roller from the number of empty slots of the left main roller (B) to obtain the number of misaligned slots of the left and right main rollers. That is, the number of misaligned slots of the left and right main rollers is 3-1=2.

[0054] 4. Assuming the slot spacing is m, where m is 185-192 micrometers, and using the misaligned slot number 2 calculated in step 3, the theoretical misaligned slot distance is: (BA)*m=2m;

[0055] 5. Install the mobile phone on the guide slide of the mobile detection terminal, and install the end plates of the long plate connected to the guide slide on the outermost side posts of the left and right main rollers respectively. Then slide the guide slide. During this process, use the macro lens on the mobile phone to photograph the misaligned end faces of the left and right main rollers, and then use image calibration and processing methods to identify feature points C, D, and E (e.g., Figure 3 As shown in the figure, the distance between the two feature points C and E is finally calculated using the pixel principle, which is the actual distance CE of the misalignment.

[0056] For example: Assuming the main roller length is H, the remaining unmachined surface at both ends is equal to I, the groove spacing is m, the groove width is J, the groove platform is K, and the number of grooves is L, then we have:

[0057] H = 2I + m * L, m = J + K

[0058] In theory, such as Figure 3 As shown; CE = (BA) * (J + K), which is the part in brackets within the dashed lines in the figure;

[0059] 6. Compare the theoretical misalignment distance calculated in step 4 with the actual misalignment distance calculated in step 5 to determine whether the installation status of the main roller and the grooving method of the diamond wire are accurate.

[0060] As in the example above, when CE = (BA) * m (i.e., CE / m = BA = number of misgrooves required by the process), it is determined that the main roller is installed correctly, the diamond wire is laid correctly, and the installation status of the main roller and the grooving method of the diamond wire are both accurate.

[0061] When CE > (BA) * m (i.e. CE / m > BA), it is considered that the main roller is not installed correctly and the installation status is inaccurate. The installation position of the left main roller is closer to the negative X-axis direction than the right main roller. At the same time, if BA = the number of misgrooves required by the process, it is considered that the diamond wire wiring is correct and the grooving method is accurate. Otherwise, it is considered that the diamond wire wiring is incorrect and the grooving method is inaccurate.

[0062] When CE < (BA) * m (i.e., CE / m < BA), it is considered that the main roller is not installed correctly and the installation status is inaccurate. The installation position of the left main roller is closer to the positive X-axis direction than the right main roller. At the same time, if BA = the number of misaligned grooves required by the process, the diamond wire wiring is correct and the grooving method is accurate. Otherwise, it is considered that the diamond wire wiring is incorrect and the grooving method is inaccurate.

[0063] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for detecting the main roller and diamond wire of a silicon wafer cutting equipment, characterized in that, include: S1. Obtain images of the grooves at the front ends of the two main rollers in the silicon wafer cutting equipment; S2. Identify the number of empty slots for the two main rollers from the acquired groove image; S3. Calculate the number of misaligned slots based on the number of empty slots on the two main rollers. The number of misaligned slots is BA, where A and B are the number of empty slots on the two main rollers, respectively. S4. Based on the number of misaligned slots and the slot spacing, calculate the theoretical distance of the misaligned slots using the following formula. (BA) × m; In the formula, A and B are the number of empty slots on the two main rollers, respectively, and m is the slot spacing. S5. Move to acquire images of the misaligned end faces of the two main rollers, then use image calibration and processing methods to identify feature points from the misaligned end face images, and then use the pixel principle to calculate the distance between the two feature points, which is the actual misalignment distance. S6. Compare the actual misalignment distance with the theoretical misalignment distance to determine whether the installation status of the main roller and the grooving method of the diamond wire are accurate; when the actual misalignment distance is equal to the theoretical misalignment distance, it is determined that the installation status of the left and right main rollers and the grooving method of the diamond wire are accurate and without abnormalities. Conversely, if the actual misalignment distance is greater than or less than the theoretical misalignment distance, it is determined that the installation status of the left and right main rollers and the grooving method of the diamond wire are inaccurate or abnormal.

2. The method according to claim 1, characterized in that, Step S2 includes: The method of edge detection, edge probing, contour drawing, polygon mapping, and region segmentation is used to comprehensively process the groove image; Identification of calibrated objects through feature values; The number of slots with diamond wires and slots without diamond wires in the slot image is calculated by data processing, thus obtaining the number of empty slots for the two main rollers.

3. The method according to claim 1, characterized in that, In step S5, a mobile detection terminal is used to move and acquire images of the misaligned end faces of the two main rollers. The mobile detection terminal includes a camera with a macro lens, a guide slide, and a long plate. The long plate is mounted on the pins on the outermost sides of the two main rollers via baffles at both ends. The guide slide is slidably mounted on the long plate. The camera is mounted on the guide slide and moves with the sliding guide slide to capture images of the misaligned end faces of the two main rollers.

4. A mobile inspection terminal for the main roller and diamond wire inspection method of the silicon wafer cutting equipment according to any one of claims 1-3, characterized in that, The mobile detection terminal is used to acquire images of the misaligned end faces of the two main rollers. It includes a camera with a macro lens, a guide slide, and a long plate. The long plate has baffles at both ends for mounting on the outermost side posts of the two main rollers of the silicon wafer cutting equipment. The guide slide is slidably mounted on the long plate. The camera is mounted on the guide slide and moves with the sliding guide slide to capture images of the misaligned end faces of the two main rollers.

5. The mobile detection terminal according to claim 4, characterized in that, The long plate is provided with a guide rail along its length, and the guide rail is perpendicular to the axis of the two main rollers. The guide slide is slidably assembled and connected to the guide rail. The macro lens of the photography device mounted on the guide slide faces downward and is located above the misaligned end of the two main rollers.

Images