Edge detection device with alignment function

Abstract

The application discloses a kind of edge detection equipment with alignment function, belong to semiconductor manufacturing equipment field, equipment includes edge detection module, its righting mechanism contains suction disc, rotating motor, two-dimensional motion module and edge finding camera, two-dimensional motion module is composed of X-axis guide rail, Y-axis guide rail, slide plate and motor, realize wafer translation and rotation alignment;Optical assembly contains camera, prism assembly, transverse motion module and camera fine adjustment mechanism, the upper prism of prism assembly, lower prism captures the light on the upper and lower edge of wafer, is integrated to form the composite image containing multiple site information by 90 degree turning prism, equipment is also provided with code reading mechanism and intelligent control unit, can be completed simultaneously wafer alignment, edge detection and identification code reading, improve detection precision and efficiency, solve the problem that additional configuration independent detection equipment is needed in the mode of "alignment and detection separation" in existing semiconductor manufacturing, wafer processing time is long, multiple transmission is easily damaged and yield is low.

Description

Technical Field

[0001] This invention belongs to the field of semiconductor manufacturing equipment technology, specifically relating to an edge detection device with alignment function. It is mainly used in the wafer preprocessing stage and can realize wafer alignment and edge quality detection during semiconductor chip production. It is adapted to the wafer preprocessing requirements before key processes such as photolithography, etching, and deposition, and is suitable for the precision and efficiency processing scenarios of large-scale semiconductor wafer production lines. Background Technology

[0002] In the semiconductor chip manufacturing process, wafer transfer and alignment are crucial preliminary steps to ensure the precision of subsequent key processes such as photolithography, etching, and deposition. The wafer cassette, the core hub connecting the wafer cassette and process equipment, has an internal alignment system responsible for the initial positioning and orientation correction of the wafers removed from the cassette. This ensures that the wafers are delivered to the subsequent process equipment stage at precise positions and angles, preventing process failures or reduced product yields due to wafer positioning deviations.

[0003] Current mainstream alignment systems typically employ mechanical or optical positioning methods to achieve wafer alignment. Mechanical positioning uses mechanical structures such as positioning pins and stops to contact the wafer edge, pushing the wafer to a preset reference position to complete alignment. Optical positioning uses optical devices such as cameras and laser sensors to collect feature information such as wafer edges, notches, and flat edges. After the algorithm calculates the deviation, it drives the mechanical structure to adjust the wafer posture to achieve alignment.

[0004] In semiconductor manufacturing, the quality of wafer edges has a significant impact on subsequent processes and product quality. Cracks, gaps, and other defects at the edges can lead to particle contamination, misalignment of photolithography patterns, and even wafer breakage. However, current technologies require separate edge inspection equipment and employ an "align first, inspect later" approach, which increases equipment costs and space requirements, prolongs processing time, and raises transmission risks. Therefore, designing a system that can simultaneously perform alignment and edge inspection has become an urgent problem to be solved. Summary of the Invention

[0005] The purpose of this invention is to provide an edge detection device with alignment function to solve the problems of the existing semiconductor manufacturing "alignment and detection separation" mode, which requires additional independent detection equipment, has long wafer processing time, is prone to damage from multiple transmissions and has a low yield.

[0006] The first objective of this invention is to provide an edge detection device with alignment function, including an edge detection module, the edge detection module including an alignment mechanism for carrying a wafer and an optical component for imaging the edge of the wafer;

[0007] The corrective mechanism includes:

[0008] Suction cups are used to hold and carry wafers.

[0009] A rotary motor, whose output shaft is connected to the suction cup, is used to drive the suction cup to rotate;

[0010] A two-dimensional motion module, whose motion platform is fixedly connected to a rotary motor, is used to drive the rotary motor and the suction cup to move in the horizontal plane;

[0011] The edge-finding camera's optical axis points towards the edge of the wafer on the chuck.

[0012] The two-dimensional motion module includes:

[0013] Fixed X-axis guide rail;

[0014] An X-axis slide plate that is slidably mounted on an X-axis guide rail and driven by an X-axis module motor;

[0015] A Y-axis guide rail is fixedly mounted on the X-axis slide plate, and the extension direction of the Y-axis guide rail is perpendicular to the X-axis guide rail.

[0016] A Y-axis slide plate that is slidably mounted on a Y-axis guide rail and driven by a Y-axis module motor;

[0017] The rotary motor is fixedly mounted on the Y-axis slide plate;

[0018] The optical components include a light source, camera, prism assembly, lateral motion module, and camera fine-tuning mechanism;

[0019] The light source illuminates the edge area of ​​the wafer in an oblique incident manner to achieve dark field illumination or low-angle illumination;

[0020] The lateral motion module is used to drive the entire imaging mechanism to move in a direction parallel to the radial direction of the wafer.

[0021] The camera fine-tuning mechanism is used to make multi-dimensional and precise adjustments to the camera;

[0022] The prism assembly includes an upper prism and a lower prism, which are respectively positioned above and below the edge of the wafer to capture light from the upper edge surface and the lower edge surface of the wafer, respectively.

[0023] The prism assembly also includes a 90-degree turning prism, which is positioned behind the optical paths of the upper and lower prisms to turn and guide the optical paths from the upper and lower prisms to the camera.

[0024] The equipment also includes a code reading mechanism, which is located near the alignment mechanism and is used to read the wafer identification code.

[0025] The second objective of this invention is to provide an edge detection method with alignment functionality, comprising the following steps:

[0026] S1 Wafer Loading and Initial Fixing: The robot transfers the wafer and places it on the suction cup of the alignment mechanism, and then vacuum adsorption is activated to fix the wafer.

[0027] S2 wafer precision alignment steps: Control the rotary motor to drive the chuck to rotate, and at the same time, use the edge-finding camera to acquire images of the wafer edge. Calculate the pose deviation of the wafer center based on the images, and drive the two-dimensional motion module to move to correct the deviation.

[0028] S3 Identification Steps: Control the wafer to rotate to a preset angle so that the identification code area on the edge of the wafer is aligned with the reader, triggering the reader to perform image acquisition and optical character recognition (OCR) on the identification code to obtain the wafer's unique serial number;

[0029] S4 Optical Path Self-Calibration Steps: The camera images a fixed reference mark set in the optical path of the prism assembly. Based on the deviation between the current image pose and the standard pose of the reference mark, a compensation signal is generated. According to the compensation signal, the camera's physical pose is adjusted or the subsequent acquired wafer image is compensated by software coordinate transformation.

[0030] S5 Detection Distance Adjustment: Based on the specifications of the wafer being inspected, the corresponding preset parameters are called, and the lateral motion module is driven to move according to the preset parameters to adjust the object distance between the imaging mechanism and the edge of the wafer to the optimal imaging position.

[0031] S6 edge imaging steps: Drive the wafer to rotate, and at the same time guide the light from multiple surfaces of the wafer edge to the camera for imaging through the prism assembly in the optical component to obtain a composite image of the wafer edge;

[0032] S7 Defect Analysis Steps: Process the composite image, segment the single composite image into sub-image regions corresponding to the wafer front chamfer, edge face, circumferential face, back face, and back chamfer, and perform defect analysis on each region.

[0033] S8 Unloading and Data Archiving: Based on the defect analysis results, the wafer is judged as qualified or unqualified; after vacuum release, the robot picks up the wafer and sorts it, and the test data is archived and saved.

[0034] Furthermore, the S6 edge imaging step includes the S6 adaptive sampling process: during the wafer rotation scanning process, the acquired images are analyzed in real time; when no defect risk features are identified in a continuous area, the first sampling mode is used for detection; when defect risk features are identified, the process switches to the second sampling mode to detect the current risk area, and the sampling density of the second sampling mode is higher than that of the first sampling mode.

[0035] Compared with the prior art, the wafer clamping mechanism provided by the present invention has the following significant and beneficial technical effects:

[0036] (1) The present invention realizes the integration of alignment and inspection. The optical inspection component is integrated into the alignment system, eliminating the need for additional independent edge inspection equipment. Alignment and edge inspection are completed simultaneously, which reduces the procurement cost of semiconductor production line equipment and reduces the space occupation of cleanrooms. It solves the problems of high cost and large footprint of traditional discrete equipment, helps the production line to be laid out compactly, and improves space utilization.

[0037] (2) Wafers do not need to be transferred between alignment and inspection equipment, the preprocessing cycle is greatly shortened, the production cycle is accelerated, and the needs of large-scale mass production are met. The total processing time for alignment and edge inspection is shortened compared with the "align first, then inspect" mode, which improves semiconductor production efficiency and breaks through the efficiency bottleneck of traditional processes.

[0038] (3) This invention reduces the number of wafer transfers, reduces the frequency of contact with the transfer mechanism, reduces friction particle contamination, avoids scratches and breakage caused by multiple transfers, reduces the damage rate in the pre-processing stage, and improves product yield; moreover, the optical detection component is equipped with a high-resolution camera and deep learning algorithm, which can accurately identify defects and has high precision in closed-loop control of the alignment mechanism, ensuring the reliability and stability of the system and laying the foundation for subsequent processes. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a structural diagram of an edge detection device with alignment function according to an embodiment of the present invention;

[0041] Figure 2 This is a front view of an edge detection device with alignment function according to an embodiment of the present invention;

[0042] Figure 3 This is a side view of an edge detection device with alignment function according to an embodiment of the present invention;

[0043] Figure 4 This is a top view of an edge detection device with alignment function according to an embodiment of the present invention.

[0044] Explanation of reference numerals in the attached diagram: 5.1-Alignment mechanism; 5.2-Code reading mechanism; 5.3-Horizontal motion module; 5.4-Adapter plate; 5.5-Upper prism; 5.6-Lower prism; 5.7-Light source; 5.8-90-degree turning prism; 5.9-Camera; 5.10-Camera fine-tuning mechanism; 5.11-Edge-finding camera. Detailed Implementation

[0045] In the following description, numerous specific details are set forth in order to provide a more thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention can be practiced without one or more of these details. In other instances, certain technical features well-known in the art have not been described in order to avoid obscuring the invention.

[0046] To fully understand this invention, detailed steps and structures will be presented in the following description to illustrate the technical solution of this invention. Preferred embodiments of the invention are described in detail below; however, in addition to these detailed descriptions, the invention may have other embodiments.

[0047] Example 1

[0048] like Figures 1 to 4 In one embodiment of the edge detection device with alignment function of the present invention, the core innovation of the edge detection device with alignment function lies in the highly integrated edge detection module. This edge detection module deeply integrates a high-precision mechanical alignment mechanism, a reliable identification system, and an innovative single-camera multi-surface optical imaging technology, and introduces intelligent self-calibration and adaptive detection algorithms, thereby reducing the complexity, cost, and maintenance difficulty of the system while ensuring high detection accuracy.

[0049] The alignment mechanism 5.1 is the starting point of the entire inspection process, and its positioning accuracy directly determines the accuracy and repeatability of subsequent edge imaging and defect detection. This mechanism is not simply a wafer carrier, but a complex system integrating precision mechanical motion, machine vision, and real-time control technologies. The alignment mechanism 5.1 includes a suction cup, a rotary motor, a two-dimensional motion module, and an edge-finding camera 5.11.

[0050] The chuck is the direct carrier of the wafer. Its top surface typically has a complex network of vacuum adsorption channels. When the wafer is placed on the chuck, the vacuum system activates, generating negative pressure through these channels to firmly and stably adhere the wafer to the chuck surface. The chuck body is usually made of rigid, thermally stable materials such as anodized aluminum alloy or stainless steel to ensure minimal deformation under stress and changes in ambient temperature. The theoretical center of rotation of the chuck, i.e., its geometric center, serves as the reference for subsequent alignment operations.

[0051] A rotary motor is the power source driving the rotation of the chuck and wafer. In a preferred embodiment of the present invention, the rotary motor is preferably a torque motor. Compared with the traditional "servo motor + reducer" solution, the torque motor adopts a direct drive method, with its rotor directly connected to the main shaft of the chuck, eliminating intermediate transmission links such as gears and belts, thus bringing multiple advantages: it can achieve high motion control precision and positioning accuracy without backlash; it can maintain stable rotation at extremely low speeds, effectively avoiding crawling, which is crucial for the edge-finding camera 5.11 to acquire clear continuous images; its high rigidity and low vibration characteristics stem from the direct drive avoiding elastic deformation of the transmission chain, thereby ensuring the stability of the imaging system; at the same time, due to the absence of mechanical contact and wear, high cleanliness is achieved, with no wear particles generated, meeting the cleanliness requirements of semiconductor manufacturing. In addition, the rotation angle of the motor is fed back to the control system in real time by a high-precision encoder, thereby achieving precise speed and position control.

[0052] The 2D motion module precisely adjusts the positions of the chuck and wafer in the horizontal plane (i.e., the XY plane) to compensate for center deviations caused during initial wafer placement. It is essentially a high-precision cross slide. The bottom layer is a fixed X-axis guide rail, securely mounted on the equipment's substrate by a bracket. The X-axis slide rests on the X-axis guide rail via a high-precision linear slider. The X-axis module motor drives a ball screw via a coupling, or converts rotational motion into linear motion of the X-axis slide along the X-axis guide rail. Above the X-axis slide, a Y-axis guide rail is fixedly mounted, extending strictly perpendicular to the X-axis. The Y-axis slide is also mounted on the Y-axis guide rail via a slider and is driven by the Y-axis module motor to move along the Y-axis. A rotary motor is fixedly mounted on the Y-axis slide. By controlling the coordinated movement of the X-axis and Y-axis module motors, the rotary motor, chuck, and wafer on it can achieve precise movement along any trajectory within the XY plane. The precision grades of the guide rails and lead screws, as well as the control resolution of the motor, together determine the final positioning accuracy of the centering system, which can typically reach the micrometer level or even the sub-micrometer level.

[0053] The Edge Finder 5.11 is a high-resolution digital camera equipped with a telecentric lens for non-contact measurement of wafer edge positions, ensuring perspective-free imaging. The Edge Finder 5.11 is precisely mounted so that its optical axis is perpendicular to the tangent direction of the wafer edge and focused on the edge region. Its field of view covers the wafer edge and captures edge images even with maximum expected deviations at its center.

[0054] The alignment process is as follows: The wafer is transferred upstream by a robotic arm or automated guide rail and initially placed onto the suction cup of the alignment mechanism 5.1. The vacuum system is activated to hold and fix the wafer in place. At this point, the geometric center of the wafer typically deviates from the image center of the edge-finding camera 5.11 by hundreds of micrometers or even more. The control system controls the rotary motor to drive the wafer to rotate at a constant and low speed. Simultaneously, the edge-finding camera 5.11 continuously captures images of the wafer edge at a fixed frequency. The image processing unit runs an edge extraction algorithm in real time, accurately locating the wafer edge in the camera coordinate system at the sub-pixel level from each frame of the image. Since the wafer is rotating, the trajectory of a point on its edge in the field of view of the edge-finding camera 5.11 should theoretically be a perfect circle. If there is a deviation in the wafer center, the actual center of this trajectory circle will deviate from the image center. By acquiring the coordinates of edge points in multiple frames of images during the rotation process, at least three non-collinear points, and using curve fitting algorithms such as the least squares method, the center coordinates (Cx, Cy) and radius R of this trajectory circle can be calculated with high precision. The offset (Δx, Δy) of the center coordinates (Cx, Cy) relative to the image center (0, 0), after being converted by the calibration parameters of the edge-finding camera 5.11, becomes the XY-direction deviation of the wafer's geometric center relative to the image center of the edge-finding camera 5.11 in the device coordinate system. The control system uses the calculated deviation (Δx, Δy) as a command to drive the X-axis and Y-axis module motors to move the corresponding distance. The two-dimensional motion module drives the suction cup and the wafer as a whole to move, aligning the wafer's geometric center with the image center of the edge-finding camera 5.11. After alignment, the wafer can be slowly rotated one full circle again, and the edge-finding camera 5.11 verifies whether the edge trajectory is now a circle centered at the image center, ensuring that the alignment accuracy meets the requirements.

[0055] After alignment is completed, or when the wafer is rotated to a specific angle, the reading mechanism 5.2 begins operation. Its task is to read the unique identification code, typically an alphanumeric sequence, laser-etched at a specific location on the wafer edge. The reading mechanism 5.2 preferably employs an OCR (Optical Character Recognition) reader. This reader typically integrates a specific wavelength illumination source and a small, high-resolution camera. The illumination source illuminates the identification code area at an appropriate angle and brightness to avoid strong glare. After the camera captures a clear character image, the built-in OCR algorithm performs image preprocessing, character segmentation, and recognition, ultimately converting the image into a text-formatted serial number. The identified serial number is uploaded to the upper-level manufacturing execution system or a local database and bound to all image data and defect information to be generated from this edge detection. This ensures end-to-end traceability of the production process; if a problem is discovered in subsequent processes, the edge detection results for a specific wafer can be quickly and accurately traced back.

[0056] After centering and code reading are completed, the core edge optical inspection stage begins. The optical components are responsible for imaging the wafer edge and mainly include a light source (5.7), a camera (5.9), a prism assembly, a lateral motion module (5.3), and a camera fine-tuning mechanism (5.10).

[0057] The lateral motion module 5.3 is a high-precision motorized linear module whose axis of motion is precisely adjusted to be parallel to the radial direction of the wafer. The entire imaging mechanism is mounted on the slider of this module via a robust adapter plate 5.4. By controlling the module's movement through a program, the imaging mechanism can adapt to the inspection requirements of wafers of different diameters and finely adjust the object distance to ensure that the wafer edge is always within the optimal depth of field of the camera 5.9, obtaining the clearest image.

[0058] To achieve high-quality imaging, this invention places a light source 5.7 near the wafer edge. This light source is preferably an LED light source with adjustable brightness and uniform emission, whose emitted light is precisely guided to the wafer edge region. The preferred illumination method is dark-field illumination or low-angle illumination, where the light illuminates the edge at a large tilt angle, preventing light reflected from the smooth surface from entering the camera, while defects (such as chipped edges, cracks, and contamination) scatter light into the camera 5.9, thus forming a bright defect image on a dark background. This improves the signal-to-noise ratio and contrast of the defects, which is a prerequisite for obtaining clear, high-contrast composite images. Together with the unique prism combination, this constitutes the optical path system for achieving single-camera multi-surface imaging.

[0059] After equipment assembly or maintenance, the pose of camera 5.9 needs to be finely calibrated using camera fine-tuning mechanism 5.10. This invention preferably employs a six-dimensional adjustment frame. This mechanism provides six degrees of freedom for fine-tuning: translation along the X, Y, and Z axes, and rotation about the X, Y, and Z axes. Through precision adjustment screws, the position and angle of camera 5.9 can be precisely adjusted to ensure that the camera sensor plane is perpendicular to the optical path and that the optical axis coincides with the optical axis of the entire prism system. This is fundamental to obtaining distortion-free, high-contrast images.

[0060] The wafer edge region to be inspected typically includes five surfaces: the top front chamfer, the front edge face, the outermost circumferential face, the back edge face, and the bottom back chamfer. Traditionally, 3-5 independent cameras and optical path systems are required to complete the inspection. The optical components of this invention achieve ingenious folding and synthesis of the optical path through their unique prism combination, and can be completed with only one camera and one set of prisms.

[0061] The upper prism 5.5 is positioned above the wafer edge. Its specific prism angle and coating design enable it to effectively capture scattered or reflected light from the upper edge surface of the wafer, primarily including the front chamfer and part of the edge facade, and refract this light into a specific outgoing direction. Similarly, the lower prism 5.6 is positioned below the wafer edge, responsible for capturing and refracting light from the lower edge surface, primarily including the back chamfer and part of the edge facade. Through precise optical design, the upper prism 5.5 and lower prism 5.6 guide the captured, originally spatially separated light paths onto a common, virtual imaging plane; that is, the information from the upper and lower edges of the wafer is optically "stitched together." The combined light path from the upper prism 5.5 and lower prism 5.6 propagates horizontally. To accommodate the typically vertically mounted layout of the camera 5.9, the light path needs to be deflected by 90 degrees. The 90-degree deflection prism 5.8 utilizes the principle of total internal reflection to precisely deflect horizontally incident light rays by 90 degrees, transforming them into rays that propagate in a vertical direction.

[0062] This transition not only changes the direction of the optical path, but more importantly, it further integrates the optical paths from the upper prism 5.5, the lower prism 5.6, and the outermost circumferential surface, guiding them to a single imaging element—camera 5.9. A suitable imaging lens is mounted in front of camera 5.9. After the aforementioned complex optical path folding and synthesis, the light information from the five surfaces at the wafer edge finally converges on the sensor target surface of camera 5.9, forming a single, composite image containing information from all the surfaces to be inspected. In the obtained composite image, different edge regions typically appear as bright or dark bands at different locations in the image. Through image processing algorithms, sub-images corresponding to different regions such as the front chamfer, edge facade, and circumferential surface can be automatically and accurately separated from this image, and defect analysis can be performed separately. This invention accomplishes the work that traditionally requires 3-5 independent cameras and optical path systems using only one camera and one set of prisms, simplifying the mechanical structure, reducing hardware, calibration, and maintenance costs, while avoiding the difficulties of synchronization and calibration between multiple cameras, significantly improving system reliability and consistency.

[0063] During long-term operation, due to the release of mechanical stress, fluctuations in ambient temperature, or slight vibrations, the relative positions of optical components may drift slowly at the micrometer level, leading to a decrease in imaging quality, changes in the detection benchmark, and ultimately affecting the accuracy and consistency of the detection results.

[0064] In some preferred embodiments, to further improve the long-term stability of the device, the present invention also integrates an intelligent control unit. The intelligent control unit includes an online automatic optical path compensation system. This system includes a tiny, known-shape fixed reference mark positioned in the optical path of the prism assembly. This fixed reference mark is located in a stable and immovable position in the optical path, such as being etched or attached to a non-light-transmitting surface of the 90-degree turning prism 5.8. The system can be configured to automatically execute a calibration process upon daily power-on, at fixed intervals, or after each batch of wafers is inspected. In calibration mode, the camera 5.9 images the reference mark. An image analysis algorithm accurately calculates the center coordinates, rotation angle, and other pose parameters of the mark in the current image and compares them with the "standard pose" stored during the initial system calibration to calculate the deviation, including translational and rotational deviations in the X and Y directions, to determine whether the optical path has shifted.

[0065] The compensation execution unit generates a compensation signal based on the calculated deviation. This compensation can be achieved in two ways: active compensation, which involves driving the camera fine-tuning mechanism 5.10 to perform physical calibration and adjust the pose of the camera 5.9; and passive compensation, which involves performing software coordinate transformation compensation on all subsequently acquired wafer edge images to correct the image position. In active compensation, the compensation signal is sent to the camera fine-tuning mechanism 5.10, which performs micro-motion according to the instruction, adjusting the spatial pose of the camera 5.9 in the opposite direction to offset the detected optical path deviation, restoring the reference marker image to the standard pose. This method corrects errors at their source. Passive compensation is used in scenarios where frequent physical adjustments are inconvenient or where the mechanism's cost is limited. The system converts the calculated deviation into a coordinate transformation matrix. Before performing defect analysis on all subsequently acquired wafer edge images, this transformation matrix is ​​applied to perform geometric correction on the image, essentially "correcting" the image to the correct position at the digital level. This method is low-cost and fast, but does not change the actual optical path.

[0066] Performing uniform high-resolution, high-frame-rate scanning of the entire circumference of a wafer generates massive amounts of data, which is time-consuming and affects the overall inspection throughput. However, defects at the wafer edge are usually not uniformly distributed, and most areas are intact.

[0067] In some preferred embodiments, to further improve the detection efficiency of the device, the intelligent control unit of the present invention includes an adaptive sampling detection module. This adaptive sampling detection module is used to optimize detection efficiency. During wafer rotation scanning, the adaptive sampling detection module analyzes the acquired images in real time. If no defect features are identified in a continuous area, the camera 5.9 and / or the rotary motor are controlled to enter a low-sampling-rate "fast scan" mode. Once the image analysis algorithm identifies potential defect risk features, such as abrupt changes in local image contrast, abnormalities in specific texture patterns, or pixel clusters exceeding a preset grayscale threshold, the module immediately triggers a switch to a high-sampling-rate "fine scan" mode to perform a detailed inspection of the risk area. After the risk area has been scanned, the system returns to the "fast scan" mode.

[0068] At the start of detection, the system scans at a high sampling rate to ensure that even the slightest defects are captured. Image processing algorithms analyze the latest acquired continuous image sequence in real time to extract features. Intact edge regions typically exhibit uniform grayscale or specific texture patterns. Abrupt changes in local image contrast are caused by obvious bright and dark edges at chipped or cracked edges; anomalies in specific texture patterns are caused by contamination that may disrupt texture uniformity; and pixel clusters exceeding a preset grayscale threshold are caused by abnormal reflections or deposits that can form abnormal bright or dark spots.

[0069] When continuous analysis of a certain length of circular area, such as corresponding to a 90-degree rotation angle, reveals no of the aforementioned risk characteristics, the system classifies the area as "low-risk." Subsequently, the adaptive sampling detection module automatically switches to "fast scan" mode. In this mode, the area can be quickly traversed by reducing the camera's 5.9 sampling frame rate and / or increasing the wafer rotation speed. Once real-time image analysis identifies any risk characteristics, the adaptive sampling detection module immediately triggers a switch back to "fine scan" mode. The camera's 5.9 frame rate returns to a high level, and the wafer rotation speed is reduced or even paused briefly to ensure the most thorough and clear imaging of the suspicious area. The algorithm performs intensive sampling and in-depth analysis of this area. To ensure no missed detections, the mode switching trigger algorithm sets a certain look-ahead and lag area to ensure safety before entering "fast scan" and to continue scanning a short distance in "fine scan" mode after leaving the risk area. This adaptive sampling strategy, without sacrificing defect detection rate, reduces the collection and processing of redundant data, improving overall detection efficiency.

[0070] Example 2

[0071] Based on the above-described device embodiments, the present invention also provides an edge detection method with alignment functionality. This method, through programmed control, coordinates and mobilizes various modules of the device to achieve a fully automated process for wafer edge detection, from automatic loading, precise alignment, and identification to high-speed, high-precision defect detection. This method embodiment not only covers basic operating steps but also incorporates the advanced functions of the aforementioned intelligent control unit, demonstrating the method's intelligence and efficiency.

[0072] An edge detection method with alignment functionality, performed using the edge detection device with alignment functionality as described above, includes the following steps:

[0073] S1 Wafer Loading and Initial Fixing: The wafer to be inspected is taken from the upstream station or carrier by an industrial robot or automated material handling system and placed stably in the center area of ​​the chuck of the alignment mechanism 5.1. After placement, the control system triggers the vacuum generator to create negative pressure on the chuck, firmly adhering the wafer to the chuck surface, thus completing the initial fixing.

[0074] S2 Wafer Precision Alignment Steps: The control system instructs the rotary motor to drive the chuck and wafer to rotate at a preset, low, and stable speed. Simultaneously, the edge-finding camera 5.11 is triggered to continuously capture images of the wafer edges passing through its field of view at a high frame rate. The image processing unit processes the acquired image sequence in real time. First, an image filtering algorithm is used to reduce noise. Then, an edge detection operator is used to accurately extract the position coordinates of the wafer edges in each frame at the sub-pixel level. Since the wafer is rotating, the trajectory of its edge points in the edge-finding camera 5.11 coordinate system should be a circle. Using curve fitting algorithms such as least squares, a circle is fitted to the edge point set extracted from multiple consecutive frames to calculate the actual center coordinates (Cx, Cy) and radius R of the current wafer edge trajectory. This actual center is compared with the image center, and after parameter conversion by the edge-finding camera 5.11 calibration, the deviation (ΔX, ΔY) of the wafer's geometric center relative to the image center in the device's XY coordinate system can be obtained. The control system uses the calculated deviation (ΔX, ΔY) as motion commands to drive the X-axis and Y-axis motors of the two-dimensional motion module to move in tandem, moving the chuck and wafer in the horizontal plane to precisely compensate for the center deviation. After the movement is completed, a verification scan is performed to confirm that the deviation has been eliminated and to ensure that the geometric center of the wafer coincides with the image center of the edge-finding camera 5.11.

[0075] S3 Identification Steps: After alignment, the control system controls a rotary motor to rotate the wafer to a preset angle, ensuring the identification code area etched on the wafer edge is directly facing the reading mechanism 5.2, preferably an OCR reader. The OCR reader triggers an illumination source and takes a picture, using its built-in OCR algorithm to recognize characters in the image and obtain the wafer's unique serial number. This serial number is immediately uploaded and linked to all subsequent inspection data for this wafer, stored in a database, enabling end-to-end quality traceability.

[0076] S4 Optical Path Self-Calibration Step: The control system moves the imaging mechanism to a "Home" position away from the wafer via the lateral motion module 5.3, aligning the camera's field of view with the fixed reference mark. The camera 5.9 images the reference mark, and the image analysis algorithm calculates parameters such as the center position and angle of the current mark image, comparing it with the initially calibrated "standard pose" to determine the optical path deviation. A compensation signal is generated based on the deviation, driving the camera fine-tuning mechanism 5.10 to directly adjust the camera pose, physically correcting the optical path; alternatively, the calculated deviation is stored as image coordinate transformation parameters. In subsequent inspections, each frame of the acquired wafer image undergoes a corresponding geometric transformation before defect analysis. To address potential minor optical path drift during long-term equipment operation, this step is performed before or periodically before the start of each batch of wafer inspection.

[0077] S5 Detection Distance Adjustment: The control system calls preset parameters according to the wafer specifications, drives the lateral motion module to move, and moves the entire imaging mechanism radially along the wafer to adjust to the optimal imaging distance most suitable for this type of wafer, ensuring that the edge area is within the optimal depth of field range of the camera lens.

[0078] S6 Edge Imaging Step: Drive the wafer to rotate at a constant speed. The upper prism 5.5 and the lower prism 5.6 capture the light from the upper and lower edge surfaces of the wafer, respectively. After being combined by the 90-degree turning prism 5.8, the light is guided to a single camera 5.9, thereby obtaining a composite image on the sensor of camera 5.9 that includes information on the front chamfer, edge face, circumferential face, back face, and back chamfer of the wafer.

[0079] S61 Adaptive Sampling: The adaptive sampling detection module runs in real-time in the background. It quickly analyzes the latest acquired consecutive image frames, extracting key features such as local contrast, texture uniformity, and grayscale distribution. The adaptive sampling detection module has preset risk feature criteria, such as abrupt changes in local image contrast, anomalies in specific texture patterns, and the presence of pixel clusters exceeding preset grayscale thresholds. If the analysis results do not trigger any risk feature criteria within a continuously scanned area at a certain angle, the adaptive sampling detection module determines that area as "low-risk." Subsequently, the system automatically switches to "fast scan mode," quickly traversing this "clean" area by reducing the camera's 5.9 sampling frame rate and / or appropriately increasing the wafer rotation speed. This significantly reduces the amount of image data that needs to be processed and stored, improving throughput. Once real-time image analysis identifies any potential risk features, the adaptive sampling detection module immediately issues a command to instantly switch to "fine scan mode." In this mode, the system increases the camera's 5.9 sampling frame rate to its maximum and may simultaneously reduce the wafer rotation speed or even pause it briefly to ensure the most thorough and clear imaging of the suspicious area, capturing any potential minute defects. Once a detailed scan of the risk area is completed, and analysis of several consecutive frames of images indicates that the risk characteristics have disappeared, the adaptive sampling detection module will control the system to switch back to "fast scan mode" to continue efficient detection.

[0080] S7 Defect Analysis Steps: First, utilizing prior knowledge, such as the relative positions of each edge surface in the synthesized image, algorithms such as thresholding and contour extraction are used to precisely segment five sub-image regions from each frame of the synthesized image, corresponding to the wafer's front chamfer, edge face, circumferential face, back face, and back chamfer. For each sub-image region, a specialized defect detection algorithm is used for analysis. For example, for chipped edges and cracks, edge detection combined with morphological operations may be used to identify abnormal protrusions or depressions in geometric shapes. For contamination, texture analysis or speckle detection algorithms may be used to identify areas inconsistent with the background texture. Detected defects are classified according to their shape, size, contrast, and other characteristics, such as chipped edges, cracks, and contamination, and their circumferential angle position and size are recorded and bound to the wafer ID.

[0081] S8 Unloading and Data Archiving: Based on preset defect tolerance standards, such as the maximum allowable number, size, and type of defects, the edge inspection results of the entire wafer are comprehensively judged, and a "qualified" or "unqualified" conclusion is output. All raw image data, processing results, defect maps, and final judgment conclusions are archived to the database after being associated with the wafer ID. After the entire inspection process is completed, the suction cup vacuum is released. The industrial robot, according to system instructions, removes the inspected wafer from the alignment mechanism 5.1 and sorts it to different downstream bins or workstations according to the judgment results, completing the entire automated inspection cycle.

[0082] In summary, this method embodiment fully leverages the advantages of the equipment through a series of precise control and analysis steps, especially achieving a highly efficient detection process of single imaging, simultaneous detection of multiple surfaces, and separate analysis. Furthermore, stability and efficiency can be further improved through self-calibration and adaptive sampling.

[0083] The preferred embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, nor is it limited to edge detection devices with alignment functions. Devices and structures not described in detail herein should be understood as being implemented in a manner common to the art. Any person skilled in the art can make many possible variations and modifications to the technical solutions of the present invention, or modify them into equivalent embodiments, without departing from the scope of the present invention. This does not affect the essential content of the present invention. Therefore, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention are still within the protection scope of the present invention.

Claims

1. An edge detection device with alignment function, comprising an edge detection module, the edge detection module including an alignment mechanism for supporting a wafer and an optical component for imaging the edge of the wafer, characterized in that: The corrective mechanism includes: Suction cups are used to hold and carry wafers. A rotary motor, whose output shaft is connected to the suction cup, is used to drive the suction cup to rotate; A two-dimensional motion module, whose motion platform is fixedly connected to a rotary motor, is used to drive the rotary motor and the suction cup to move in the horizontal plane; The edge-finding camera's optical axis points towards the edge of the wafer on the chuck. The two-dimensional motion module includes: Fixed X-axis guide rail; An X-axis slide plate that is slidably mounted on an X-axis guide rail and driven by an X-axis module motor; A Y-axis guide rail is fixedly mounted on the X-axis slide plate, and the extension direction of the Y-axis guide rail is perpendicular to the X-axis guide rail. A Y-axis slide plate that is slidably mounted on a Y-axis guide rail and driven by a Y-axis module motor; The rotary motor is fixedly mounted on the Y-axis slide plate; The optical components include a light source, camera, prism assembly, lateral motion module, and camera fine-tuning mechanism; The light source illuminates the edge area of ​​the wafer in an oblique incident manner to achieve dark field illumination or low-angle illumination; The lateral motion module is used to drive the entire imaging mechanism to move in a direction parallel to the radial direction of the wafer. The camera fine-tuning mechanism is used to make multi-dimensional and precise adjustments to the camera; The prism assembly includes an upper prism and a lower prism, which are respectively positioned above and below the edge of the wafer to capture light from the upper edge surface and the lower edge surface of the wafer, respectively. The prism assembly also includes a 90-degree turning prism, which is positioned behind the optical paths of the upper and lower prisms to turn and guide the optical paths from the upper and lower prisms to the camera. The equipment also includes: The code reading mechanism, located near the alignment mechanism, is used to read the wafer identification code; The intelligent control unit includes an online automatic optical path compensation system and an adaptive sampling and detection module; The online automatic optical path compensation system includes a fixed reference mark set in the optical path of the prism assembly. Under predetermined conditions, the online automatic optical path compensation system controls the camera to image the fixed reference mark, calculates the pose parameters of the fixed reference mark in the current image through image analysis algorithms, compares the current pose parameters with the pre-calibrated standard pose parameters, and generates a compensation signal based on the comparison result. The compensation signal is used to drive the camera fine-tuning mechanism to perform physical pose calibration of the camera, or to perform software coordinate transformation compensation on the wafer edge image subsequently acquired by the camera. The adaptive sampling and detection module analyzes the acquired images in real time during wafer rotation scanning and dynamically controls the camera's sampling frame rate and / or the rotation speed of the rotary motor based on the identification results of defect risk features in the images. Specifically, when no defect risk features are identified within a predetermined angle area of ​​continuous scanning, the module switches to fast scanning mode; when defect risk features are identified in the image, the module immediately switches to fine scanning mode to detect the risk areas where defect risk features have been identified.

2. The edge detection device with alignment function according to claim 1, characterized in that: The rotary motor is a torque motor.

3. The edge detection device with alignment function according to claim 1, characterized in that: The reading mechanism is an OCR optical character recognition reader.

4. The edge detection device with alignment function according to claim 1, characterized in that: Defect characteristics include sudden changes in local image contrast, abnormal specific texture patterns, or pixel clusters exceeding the preset grayscale threshold.

5. The edge detection device with alignment function according to claim 1, characterized in that: The upper and lower prisms guide the captured light path to the same imaging plane, and after being integrated by a 90-degree turning prism, a composite image is formed on the camera sensor, which includes information on the wafer's front chamfer, edge face, circumferential face, back face, and back chamfer.

Images