Wafer alignment method, wafer bonding method, and wafer bonding apparatus

By introducing geometric feature points of wafer notches and reference points of arc edges into the wafer alignment method, and adopting the rigid body geometry two-point alignment principle, adaptive alignment without template matching is achieved, which solves the problems of insufficient accuracy and stability in the existing technology and improves the accuracy and anti-interference ability of wafer alignment.

CN123076301APending Publication Date: 2026-11-10SABERS CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611529825.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-30
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing wafer alignment methods suffer from low precision stability and poor environmental adaptability, failing to meet the requirements of high-precision bonding.

Method used

By adopting the principle of rigid body geometry two-point alignment, the wafer reference vector is determined by obtaining the geometric feature points of the wafer notch and the reference points of the arc edge, thus achieving adaptive alignment without template matching and eliminating the need for interference-prone Notch vertices and wafer center fitting.

Benefits of technology

It improves the accuracy and stability of wafer alignment, enhances anti-interference capabilities, and eliminates the need for template creation and iterative updates.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123076301A_ABST
    Figure CN123076301A_ABST
Patent Text Reader

Abstract

Embodiments of the present application provide wafer alignment methods, wafer bonding methods and wafer bonding devices. The method comprises: obtaining a first gap image of a first wafer; obtaining a first wafer reference vector through a wafer reference vector determination process; obtaining a second gap image of a second wafer; obtaining a second wafer reference vector through a wafer reference vector determination process; and aligning the second wafer reference vector with the first wafer reference vector to align the second wafer with the first wafer. The wafer reference vector determination process comprises: obtaining a geometric feature point of the gap as a first reference point based on the gap image; obtaining a second reference point on a predetermined position of an arc edge of one side of the wafer; and determining the wafer reference vector. Embodiments of the present application improve the wafer alignment accuracy, stability and anti-interference ability in the bonding process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of semiconductor manufacturing, and particularly relates to a wafer alignment method, a wafer bonding method, a wafer bonding apparatus, a computer-readable storage medium, and a computer program product. Background Technology

[0002] In semiconductor manufacturing, wafer bonding includes alignment and bonding. Wafer alignment comprises two core stages: pre-alignment and fine alignment. Current technologies primarily employ notch template matching and traditional geometric alignment for the alignment stage.

[0003] In the notch template matching method, a standard notch template needs to be acquired before the system runs. During the actual alignment process, after the wafer to be tested is fed into the equipment, the vision system captures images of the wafer edge in real time. Subsequently, the system compares the real-time extracted notch features with the pre-stored standard template, and adjusts the position and angle of the wafer according to the deviation data until the actual position of the notch completely coincides with the preset target position, providing a foundation for subsequent high-precision bonding. In this notch template matching method, changes in illumination, wafer edge oxidation, dust adhesion, and slight edge chipping can cause matching score drift; large-range rotation matching has a large computational load and poor real-time performance; slight differences in the shape of wafers from different batches and grayscale shifts due to equipment aging all require the template to be remade, resulting in high mass production maintenance costs and low long-term positioning repeatability accuracy.

[0004] Traditional geometric alignment methods rely on wafer-wide circular fitting to obtain the wafer center, and then use Notch vertices or contour orientation angles to construct an alignment reference. Circular fitting is susceptible to interference from wafer outer contour damage and edge impurities, introducing systematic errors. Directly using Notch vertices as feature points has weak anti-interference capabilities and exhibits significant single-point jitter. Furthermore, conventional angle calculations are prone to jumps between 0° and 180°, causing the rotating mechanism to reverse its movement and leading to alignment failure.

[0005] In summary, the alignment methods used in the existing alignment stage have low accuracy stability and poor environmental adaptability, and cannot meet the requirements of high-precision alignment.

[0006] Therefore, there is an urgent need for an adaptive alignment method with high alignment accuracy, stability, and strong anti-interference ability to provide a good foundation for subsequent high-precision bonding. Summary of the Invention

[0007] The present invention provides a wafer alignment method, a wafer bonding method, a wafer bonding apparatus, a computer-readable storage medium, and a computer program product, which can improve the wafer alignment accuracy and stability, as well as the adaptability and anti-interference capability during the bonding process.

[0008] In a first aspect, embodiments of the present invention provide a wafer alignment method, wherein the wafer has a notch for positioning. The wafer alignment method includes: Obtain the first notch image of the first notch on the first wafer; The first wafer reference vector of the first wafer is obtained by performing the wafer reference vector determination process; Obtain the second notch image of the second wafer; The second wafer reference vector of the second wafer is obtained by performing the wafer reference vector determination process. Align the second wafer reference vector with the first wafer reference vector to align the second wafer with the first wafer. The wafer reference vector determination process includes: obtaining the geometric feature points of the notch based on the notch image as a first reference point; obtaining a second reference point at a predetermined position on the arc edge of one side of the wafer, so as to determine the vector from one of the first reference point and the second reference point to the other as the wafer reference vector.

[0009] In some embodiments of the present invention, before obtaining the geometric feature points of the notch based on the notch image as the first reference point, the wafer reference vector determination process further includes: Based on the tilt angle of the notch determined from the notch image, the wafer is driven so that the notch is aligned with a standard orientation; The step of obtaining the geometric feature points of the notch as the first reference point based on the notch image includes: after driving the wafer to align the notch with the standard direction, obtaining a new notch image, and obtaining the geometric feature points of the notch aligned with the standard direction in the new notch image as the first reference point.

[0010] In some embodiments of the present invention, obtaining the first notch image of the first notch of the first wafer includes: obtaining a first original image of the first wafer acquired by a first original image acquisition device; performing adaptive image enhancement on the first original image and delineating the region of interest where the first notch is located to obtain a first processed image; performing adaptive threshold segmentation on the first processed image to extract the first notch image; The step of obtaining the second notch image of the second wafer includes: acquiring a second original image of the second wafer acquired by a first original image acquisition device; performing adaptive image enhancement on the second original image and delineating the region of interest where the second notch is located to obtain a second processed image; and performing adaptive threshold segmentation on the second processed image to extract the second notch image.

[0011] In some embodiments of the present invention, the step of performing image adaptive enhancement on the first original image and delineating the region of interest where the first gap is located to obtain a first processed image includes: enhancing the contrast of the image edge of the first gap based on the first original image to obtain a first preprocessed image; smoothing the noise in the first preprocessed image to obtain a second preprocessed image; and delineating the region of interest where the first gap is located in the second preprocessed image to obtain the first processed image. The step of performing adaptive image enhancement on the second original image and defining the region of interest where the second gap is located to obtain a second processed image includes: enhancing the contrast of the image edge of the second gap based on the second original image to obtain a third preprocessed image; smoothing the noise in the third preprocessed image to obtain a fourth preprocessed image; and defining the region of interest where the second gap is located in the fourth preprocessed image to obtain the second processed image.

[0012] In some embodiments of the present invention, the step of enhancing the contrast of the image edge of the first notch based on the first original image to obtain a first preprocessed image includes: maximizing the enhancement of the contrast of the image edge of the first notch by power transformation and grayscale stretching based on the first original image to obtain the first preprocessed image; The step of smoothing the noise in the first preprocessed image to obtain the second preprocessed image includes: smoothing the noise in the first preprocessed image using morphological closing operations to obtain the second preprocessed image; The step of enhancing the contrast of the image edge of the second notch based on the second original image to obtain a third preprocessed image includes: maximizing the enhancement of the contrast of the image edge of the second notch by power transformation and grayscale stretching based on the second original image to obtain the third preprocessed image; Smoothing the noise in the third preprocessed image to obtain the fourth preprocessed image includes: smoothing the noise in the third preprocessed image using morphological closing operations to obtain the fourth preprocessed image.

[0013] In some embodiments of the present invention, the step of performing adaptive threshold segmentation on the first processed image to extract the first gap image includes: using adaptive standard deviation local threshold segmentation to extract the dark region of the first gap from the first processed image to obtain a fifth preprocessed image; filtering out interference in the fifth preprocessed image to obtain a sixth preprocessed image; and separating the complete first target region of the first gap from the sixth preprocessed image by area filtering. The first gap image is obtained by performing morphological opening and closing operations and region differencing on the image of the first target region. The step of performing adaptive threshold segmentation on the second processed image to extract the second gap image includes: using adaptive standard deviation local threshold segmentation to extract the dark region of the second gap from the second processed image to obtain a seventh preprocessed image; filtering out interference in the seventh preprocessed image to obtain an eighth preprocessed image; separating the complete second target region image of the second gap from the eighth preprocessed image through area filtering; and performing morphological opening and closing operations and region differencing on the image of the second target region to obtain the second gap image.

[0014] In some embodiments of the present invention, driving the wafer to align the notch with a standard direction based on the tilt angle of the notch determined by the notch image includes: determining two target endpoints of the notch based on the notch image; calculating the tilt angle of the notch based on the two target endpoints; and driving the wafer to rotate based on the tilt angle to align the notch with the standard direction.

[0015] In some embodiments of the present invention, determining the two target endpoints of the notch based on the notch image includes: using Canny subpixel edge detection to extract the first contour lines on both sides of the notch in the notch image; expanding the notch in the notch image to obtain the second contour line of the expanded notch; and determining the intersection of the first contour line and the second contour line as the two target endpoints. The step of calculating the tilt angle of the gap based on the two target endpoints includes: calculating at least two parameters among the height difference between the two target endpoints, the width between the two target endpoints, and the distance between the two target endpoints; and calculating the tilt angle of the gap based on the at least two parameters.

[0016] In some embodiments of the present invention, calculating the tilt angle of the notch based on the at least two parameters includes: calculating the initial tilt angle of the notch based on the at least two parameters; when the initial tilt angle is greater than 90°, subtracting 180° from the initial tilt angle to obtain the tilt angle; when the initial tilt angle is less than -90°, adding 180° to the initial tilt angle to obtain the tilt angle; and when the initial tilt angle is greater than or equal to -90° and less than or equal to 90°, using the initial tilt angle as the tilt angle.

[0017] In some embodiments of the present invention, the first reference point is the center of the largest inscribed circle of the notch. Alternatively, the first reference point can be the center of a circle obtained by transforming the outline of the gap through the vertex of the gap. Alternatively, the first reference point can be the midpoint of the line connecting the endpoints on both sides of the notch. Alternatively, the first reference point is a position point at which the distance between the two specific lines of the gap image is a first preset distance and a second preset distance, respectively. The two specific lines are located on both sides of the vertex of the gap, and for each of the two specific lines, the specific line is the line that can pass through the target contour points on the contour of that side of the gap from the corresponding side of the vertex of the gap.

[0018] In some embodiments of the present invention, the first reference point is a position point at which the distance between it and two specific straight lines of the notch image is a first preset distance and a second preset distance, respectively. The step of obtaining the geometric feature points of the notch based on the notch image as the first reference point includes: Two target line segments are identified in the gap image, each of the two target line segments being a partial outline of the gap in the gap image; For each of the two target line segments, the straight line that passes through the most target contour points on that target line segment is determined as the specific straight line corresponding to that target line segment, so as to determine the two specific straight lines in the gap image; The point in the gap image that is at a distance of the first specific line from the two specific lines and at a distance of the second specific line from the two specific lines is determined as the geometric feature point of the gap and is used as the first reference point.

[0019] In some embodiments of the present invention, the first reference point is the center of a circle obtained by morphological transformation of the outline of the gap, passing through the vertex of the gap. The step of obtaining the geometric feature points of the gap as the first reference point based on the gap image includes: smoothing and denoising the outline of the gap in the gap image; performing contour shape transformation on the smoothed and denoised outline of the gap to extract the circle passing through the vertex of the gap; calculating the center of the circle and using the center of the circle as the first reference point.

[0020] In some embodiments of the present invention, the second reference point is the intersection of the arc edge on one side of the wafer and the predetermined axis of the original image acquired by the second original image acquisition device.

[0021] In some embodiments of the present invention, obtaining a second reference point at a predetermined position on the arc edge of one side of the wafer includes: obtaining a local arc original image acquired by the second original image acquisition device, wherein the local arc original image includes the arc edge; extracting the arc edge from the local arc original image; obtaining the intersection of the arc edge and the predetermined axis as the second reference point, wherein the predetermined axis is an axis at a predetermined position in the width direction of the local arc original image in the camera coordinate system where the second original image acquisition device is located.

[0022] In some embodiments of the present invention, the predetermined axis is the central axis in the width direction of the local circular arc original image in the camera coordinate system where the second original image acquisition device is located.

[0023] In some embodiments of the present invention, aligning the second wafer reference vector with the first wafer reference vector to align the second wafer with the first wafer includes: calculating the offset and angular deviation between the second wafer reference vector and the first wafer reference vector based on the second wafer reference vector and the first wafer reference vector; and aligning the second wafer with the first wafer based on the offset and the angular deviation.

[0024] In some embodiments of the present invention, the wafer alignment method is applied to a wafer bonding apparatus, the wafer bonding apparatus comprising: a first stage for holding one of a first wafer and a second wafer during wafer bonding, wherein the first wafer has a first arc edge and a first notch for positioning, and the second wafer has a second arc edge and a second notch for positioning; a second stage for holding the other of the first wafer and the second wafer during wafer bonding; a first original image acquisition unit for acquiring original images of the first wafer and the second wafer; and a second original image acquisition unit for acquiring a partial arc original image of the first wafer including the first arc edge and a partial arc original image of the second wafer including the second arc edge. The wafer alignment method includes: acquiring an original image of the first wafer by a first original image acquisition device to obtain a first notch image of the first notch; obtaining geometric feature points of the first notch as the first reference point of the first wafer based on the first notch image; acquiring a partial arc original image of the first wafer including the first arc edge by a second original image acquisition device to obtain a second reference point of the first wafer and determine a first wafer reference vector; acquiring an original image of the second wafer by the first original image acquisition device to obtain a second notch image of the second notch; obtaining geometric feature points of the second notch as the first reference point of the second wafer based on the second notch image; acquiring a partial arc original image of the second wafer including the second arc edge by the second original image acquisition device to obtain a second reference point of the second wafer and determine a second wafer reference vector, wherein the relative positional relationship between the first reference point and the second reference point of the first wafer is the same as the relative positional relationship between the first reference point and the second reference point of the second wafer; and aligning the second wafer reference vector with the first wafer reference vector by driving the first stage and the second stage to align the second wafer with the first wafer.

[0025] Secondly, embodiments of the present invention provide a wafer bonding method, comprising: aligning the second wafer and the first wafer according to the wafer alignment method described in any one of the foregoing claims; and bonding the first wafer and the second wafer.

[0026] Thirdly, embodiments of the present invention provide a wafer bonding apparatus, comprising: A first stage is used to hold one of a first wafer and a second wafer during wafer bonding, wherein the first wafer has a first arc edge and a first notch for positioning, and the second wafer has a second arc edge and a second notch for positioning; A second stage is used to hold one of the first wafer and the second wafer during wafer bonding; A first raw image acquisition device is used to acquire raw images of the first wafer and raw images of the second wafer; The second raw image acquisition device is used to acquire a local arc raw image of the first wafer containing the first arc edge and to acquire a local arc raw image of the second wafer containing the second arc edge. The processor and the memory storing computer program instructions; the processor reads and executes the computer program instructions to implement any of the wafer alignment methods described in the above embodiments.

[0027] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the wafer alignment method described in any one of the preceding claims.

[0028] Fifthly, embodiments of the present invention provide a computer program product, including a computer program that, when executed by a processor, implements any of the wafer alignment methods described above.

[0029] The wafer alignment method, wafer bonding method, wafer bonding device, computer-readable storage medium, and computer program product provided in this invention are based on the rigid body geometry two-point alignment principle. By introducing reference vectors of a first wafer and a second wafer, they align the second wafer with the first wafer, thus providing a template-free adaptive alignment scheme that eliminates the need for template creation, storage, and iterative updates. Furthermore, the wafer alignment method provided in this invention abandons the easily disturbed Notch vertex and wafer center fitting, improving alignment accuracy, stability, and anti-interference capability. Attached Figure Description

[0030] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 This is a schematic flowchart of the wafer alignment method provided in an embodiment of the present invention; Figure 2 As shown Figure 1 A flowchart illustrating a specific example of a wafer alignment method; Figure 3 It shows Figure 1 A flowchart illustrating another specific example of the wafer alignment method; Figure 4 This is a schematic diagram of a wafer bonding device provided in an embodiment of the present invention; Figure 5 This is a schematic flowchart of a wafer alignment method provided in other embodiments of the present invention; Figure 6 This is a schematic flowchart of a wafer alignment method provided in other embodiments of the present invention; Figure 7 This is a schematic flowchart of a wafer alignment method provided in other embodiments of the present invention; Figure 8 yes Figure 7 The illustration shows a schematic diagram of the wafer alignment method applied in a wafer bonding device. Figure 9This is a schematic flowchart of a wafer alignment method provided in other embodiments of the present invention; Figure 10 yes Figure 9 A schematic diagram of step S102 in the wafer alignment method shown; Figure 11 yes Figure 9 A schematic diagram of step S302 in the wafer alignment method shown; Figure 12 yes Figure 9 Another flowchart of step S102 in the wafer alignment method shown; Figure 13 yes Figure 9 Another flowchart illustrating step S302 in the wafer alignment method shown; Figure 14 yes Figure 9 A schematic diagram of step S103 in the wafer alignment method shown; Figure 15 yes Figure 9 A schematic diagram of step S303 in the wafer alignment method shown; Figure 16 yes Figure 3 A schematic diagram of step S203 in the wafer alignment method shown; Figure 17 yes Figure 3 A schematic diagram of step S403 in the wafer alignment method shown; Figure 18 yes Figure 16 The flowchart of steps S2031 and S2032 in the wafer alignment method shown is illustrated. Figure 19 yes Figure 17 The flowchart of steps S4031 and S4032 in the wafer alignment method shown is illustrated. Figure 20 yes Figure 18 The flowchart of step S20322 in the wafer alignment method shown is as follows; Figure 21 yes Figure 19 The flowchart of step S40322 in the wafer alignment method shown is illustrated. Figure 22 yes Figure 2 A schematic diagram of step S201 in the wafer alignment method shown; Figure 23 yes Figure 2 A schematic diagram of step S401 in the wafer alignment method shown; Figure 24 yes Figure 2 Another flowchart of step S201 in the wafer alignment method shown; Figure 25 yes Figure 2 Another flowchart illustrating step S401 in the wafer alignment method shown; Figure 26 yes Figure 2 A schematic diagram of step S202 in the wafer alignment method shown; Figure 27 yes Figure 2 A schematic diagram of step S402 in the wafer alignment method shown; Figure 28 This is a flowchart illustrating step S50 in the wafer alignment method provided in this embodiment of the invention. Figure 29 A flowchart of a wafer bonding method provided in an embodiment of the present invention; Figure 30 This is a flowchart of a wafer bonding method provided in a specific embodiment of the present invention. Detailed Implementation

[0032] The features and exemplary embodiments of various aspects of the present invention will now be described in detail. To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely intended to explain the present invention and not to limit the present invention. For those skilled in the art, the present invention can be practiced without some of these specific details. The following description of the embodiments is merely to provide a better understanding of the present invention by illustrating examples of the invention.

[0033] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0034] To address the problems of the prior art, embodiments of the present invention provide a wafer alignment method, a wafer bonding method, a wafer bonding apparatus, a computer-readable storage medium, and a computer program product.

[0035] The wafer alignment method provided in the embodiments of the present invention will be described by way of example below.

[0036] The wafer alignment method is applicable to wafers with notches for positioning. According to embodiments of the invention, the shape of the notch is not limited. For example, the notch can be a V-shaped notch or a U-shaped notch. Figure 1 This is a schematic flowchart of the wafer alignment method provided in an embodiment of the present invention. Figure 1 As shown, the wafer alignment method includes the following steps: S10: Obtain the first notch image of the first notch on the first wafer; S20: Obtain the first wafer reference vector of the first wafer by performing the wafer reference vector determination process; S30: Obtain the second notch image of the second notch on the second wafer; S40: Obtain the second wafer reference vector of the second wafer by performing the wafer reference vector determination process; S50: Align the second wafer reference vector with the first wafer reference vector to align the second wafer with the first wafer, wherein the wafer reference vector determination process includes: obtaining the geometric feature point of the notch based on the notch image as a first reference point; obtaining a second reference point at a predetermined position on the arc edge of one side of the wafer, so as to determine the vector pointing from one of the first reference point and the second reference point to the other as the wafer reference vector.

[0037] The wafer alignment method provided in this invention is based on the rigid body geometry two-point alignment principle. It aligns the second wafer with the first wafer by introducing reference vectors for the first and second wafers, thus providing a template-free adaptive alignment scheme that eliminates the need for template creation, storage, and iterative updates. Furthermore, the wafer alignment method provided in this invention abandons the easily disturbed Notch vertices and wafer center fitting, improving alignment accuracy, stability, and anti-interference capability.

[0038] Figure 2 It shows Figure 1 A schematic flowchart illustrating a specific example of a wafer alignment method. In some embodiments of the present invention, such as... Figure 2 As shown, step S20: Obtaining the first wafer reference vector of the first wafer by performing the wafer reference vector determination process, including: S201: Obtain the geometric feature points of the first gap as the first reference point based on the first gap image; S202: Obtain a second reference point at a predetermined position on the first arc edge of one side of the first wafer, so as to determine the vector of the first wafer pointing from one of the first reference point and the second reference point to the other as the first wafer reference vector.

[0039] Furthermore, such as Figure 2 As shown, step S40: Obtaining the second wafer reference vector of the second wafer by performing the wafer reference vector determination process includes: S401: Obtain the geometric feature points of the second notch based on the second notch image as the first reference point; S402: Obtain a second reference point at a predetermined position on the second arc edge of one side of the second wafer, so as to determine the vector of the second wafer pointing from one of the first reference point and the second reference point to the other as the second wafer reference vector.

[0040] It is understood that the geometric feature points of the first gap are obtained based on the image of the first gap, and the geometric feature points of the second gap are obtained based on the image of the second gap.

[0041] The first reference point of the first wafer reference vector corresponds to the first reference point of the second wafer reference vector, and the second reference point of the first wafer reference vector corresponds to the second reference point of the second wafer reference vector. That is, the first wafer reference vector is a vector pointing from the first reference point of the first wafer to the second reference point, and the second wafer reference vector is a vector pointing from the first reference point of the second wafer to the second reference point.

[0042] For example, for any one of the first and second notch images, the geometric feature points of the notch in that notch image are unique and can be uniquely determined. These geometric feature points can be determined using plane geometry principles. For instance, they can be obtained by curve fitting based on the notch contour of the notch image; or, by traversing the notch contour and calculating the curvature of each contour point, the extreme curvature points can be determined as the geometric feature points of the notch. For example, for the second reference point at a predetermined position on the arc edge, it can be determined by combining a preset axis in the image space—a scheme described in detail below. Alternatively, the wafer center can be determined in the acquired local arc image, and a corresponding point can be determined on the arc edge based on a preset angle, using this point as the second reference point.

[0043] It should be noted that the wafer reference vector determined by the first and second reference points is a virtual reference vector, and there are no corresponding actual line segments on the wafer surface. The first reference point is extracted from the notch image of the wafer; the second reference point is a point determined on the arc edge of the wafer. In this embodiment of the invention, the wafer reference vector refers to the vector formed by the first and second reference points in a unified coordinate system. More specifically, a virtual direction vector representing the wafer orientation can be obtained based on the positions of the two reference points in the same coordinate system. Subsequent wafer alignment operations essentially involve rotating the virtual reference vector of the second wafer until it coincides with the direction of the virtual reference vector of the first wafer, thereby completing the alignment of the two wafers.

[0044] As described above, the wafer alignment method provided by the embodiments of the present invention is based on the rigid body geometry two-point alignment principle. The first wafer reference vector is determined by the first reference point and the second reference point of the first wafer, and the second wafer reference vector is determined by the first reference point and the second reference point of the second wafer. Then, the second wafer and the first wafer are aligned by aligning the second wafer reference vector with the first wafer reference vector. Therefore, the wafer alignment method according to the embodiments of the present invention provides a template-free adaptive alignment scheme, which can eliminate the need for template creation, storage, and iterative updates. Furthermore, the wafer alignment method according to the embodiments of the present invention abandons the easily disturbed Notch vertices and wafer center fitting, which can improve alignment accuracy, stability, and anti-interference capability.

[0045] In some embodiments of the present invention, before obtaining the geometric feature points of the notch based on the notch image as the first reference point, the wafer reference vector determination process further includes: Based on the tilt angle of the notch determined from the notch image, the wafer is driven so that the notch is aligned with a standard orientation; The step of obtaining the geometric feature points of the notch as the first reference point based on the notch image includes: after driving the wafer to align the notch with the standard direction, obtaining a new notch image, and obtaining the geometric feature points of the notch aligned with the standard direction in the new notch image as the first reference point.

[0046] The standard direction refers to a fixed direction pointing towards the center of the wafer on the wafer plane. For example, the standard direction may be a reference direction specified by the wafer bonding equipment.

[0047] Figure 3 It shows Figure 1 A flowchart illustrating another specific example of a wafer alignment method. In some embodiments of the present invention, such as... Figure 3As shown, before step S201: obtaining the geometric feature points of the first notch as the first reference point based on the first notch image, step S20: obtaining the first wafer reference vector of the first wafer by performing the wafer reference vector determination process, further includes: S203: Based on the first tilt angle of the first notch determined by the first notch image, drive the first wafer so that the first notch is aligned with the standard direction; Step S201: Obtaining the geometric feature points of the first notch as the first reference point based on the first notch image includes: S2011: Obtain the geometric feature point of the first notch aligned with the standard direction as the first reference point.

[0048] Furthermore, such as Figure 3 As shown, before step S401: obtaining the geometric feature points of the second notch as the first reference point based on the second notch image, step S40: obtaining the second wafer reference vector of the second wafer by performing the wafer reference vector determination process, further includes: S403: Based on the second tilt angle of the second notch determined by the second notch image, drive the second wafer so that the second notch is aligned with the standard direction; Step S401: Obtaining the geometric feature points of the second notch as the first reference point based on the second notch image includes: S4011: Obtain the geometric feature point of the second notch aligned with the standard direction as the first reference point.

[0049] It is understandable that after driving the wafer to move so that the wafer notch is aligned with the standard direction, the notch image can be re-acquired for the wafer notch.

[0050] In the above embodiments of the present invention, before determining the first reference point of the wafer reference vector, the wafer is driven according to the tilt angle so that the notch is aligned with the standard direction, thereby avoiding the problem of asymmetric notch contour extraction caused by wafer tilt and improving the accuracy of wafer alignment.

[0051] In some embodiments of the present invention, the wafer alignment method is applied to a wafer bonding device. Figure 4 This is a schematic diagram of a wafer bonding apparatus provided in an embodiment of the present invention. Figure 4 As shown, the wafer bonding apparatus may include a first stage 10, a second stage 20, a first raw image acquisition unit 30, and a second raw image acquisition unit 40.

[0052] In this embodiment, the first stage 10 is used to hold the first wafer 50 during wafer bonding. (Reference) Figure 4 The first wafer 50 has a first notch 51 and a first arc edge 52 for positioning, and the second wafer 60 has a second notch 61 and a second arc edge 62 for positioning. The wafer bonding process refers to the process of bonding the second wafer 60 to the first wafer 50 after aligning the first wafer 50 and the second wafer 60 in the wafer bonding method.

[0053] The second stage 20 is used to hold the second wafer 60 during wafer bonding.

[0054] It is understood that in some other embodiments, the first stage 10 can be used to hold the second wafer 60 during wafer bonding, and the second stage 20 can be used to hold the first wafer 50 during wafer bonding.

[0055] Continue reading Figure 4 The first raw image acquisition unit 30 is used to acquire raw images of the first wafer 50 and the second wafer 60.

[0056] The second original image acquisition unit 40 is used to acquire a partial arc original image of the first wafer 50 including the first arc edge 52 and to acquire a partial arc original image of the second wafer 60 including the second arc edge 62.

[0057] For example, refer to Figure 4 The first raw image acquisition unit 30 is used to acquire raw images of the first wafer 50 held on the second stage 20 and the second wafer 60 held on the second stage 20. It is understood that each of these two raw images may contain the notch of the corresponding wafer. The second raw image acquisition unit 40 is used to acquire raw images of a partial arc of the first wafer 50 held on the second stage 20, including the first arc edge 52, and raw images of a partial arc of the second wafer 60 held on the second stage 20, including the second arc edge 62.

[0058] According to an embodiment of the present invention, for ease of explanation, the holding plane of the second stage 20 used to hold the wafer can be used as a reference plane, and the center point of the reference plane can be used as the origin to establish a reference coordinate system. This reference coordinate system can be a Cartesian coordinate system. In a Cartesian coordinate system, the X-axis and Y-axis are along the horizontal and vertical directions, respectively, and the Z-axis is along a direction perpendicular to the reference plane.

[0059] In some embodiments of the present invention, the angle between the mounting positions of the first raw image acquisition unit 30 and the second raw image acquisition unit 40 on the circumference of the wafer is a preset angle. For example, the preset angle is 90°. For instance, the first raw image acquisition unit 30 is arranged on the Y-axis of the reference coordinate system, and the second raw image acquisition unit 40 is arranged on the X-axis of the reference coordinate system. In this case, the standard direction can be the Y-axis direction.

[0060] In some embodiments of the present invention, the first platform 10 is capable of at least one of X-axis motion, Y-axis motion, Z-axis motion, and rotation about the Z-axis, while the remaining dimensions of motion are achieved by the second platform 20. That is, the four dimensions of motion—X-axis motion, Y-axis motion, Z-axis motion, and rotation about the Z-axis—can be allocated to the first platform 10 and the second platform 20 as needed.

[0061] In other embodiments of the present invention, such as Figure 4 As shown, the first platform 10 can achieve Z-axis movement, and the second platform 20 can achieve X-axis and Y-axis movement as well as rotation around the Z-axis.

[0062] In other embodiments of the present invention, the first platform 10 is capable of X-axis and Y-axis movement and rotation about the Z-axis, and the second platform 20 is capable of Z-axis movement.

[0063] In other embodiments of the present invention, the first platform 10 is capable of X-axis movement, and the second platform 20 is capable of Y-axis movement, Z-axis movement, and rotation about the Z-axis.

[0064] In some embodiments of the present invention, the first stage 10 and the second stage 20 are chucks. Exemplarily, the first stage 10 and the second stage 20 can hold the wafer by adsorption.

[0065] In some embodiments of the present invention, such as Figure 4 As shown, the control device (such as a driver) of the first stage 10, the control device (such as a driver) of the second stage 20, the first raw image acquisition unit 30, the second raw image acquisition unit 40, the processor 70, and the memory 80 are connected to each other, for example, via a bus 90.

[0066] Bus 90 includes hardware, software, or both, that couples components of a wafer bonding device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, bus 90 may include one or more buses. Although specific buses are described and illustrated in the embodiments of the present invention, the present invention is contemplated by any suitable bus or interconnect.

[0067] Figure 5 This is a schematic flowchart of a wafer alignment method provided in other embodiments of the present invention. In some embodiments of the present invention, such as... Figure 5 As shown, the wafer alignment method is applied to a wafer bonding device and includes the following steps: S100: The first original image acquisition device acquires the original image of the first wafer to obtain the first notch image of the first notch; S110: Based on the first notch image, obtain the geometric feature points of the first notch as the first reference point of the first wafer; S120: The second original image acquisition device acquires a local arc original image of the first wafer containing the first arc edge, so as to obtain the second reference point of the first wafer and determine the reference vector of the first wafer. S130: The first original image acquisition device acquires the original image of the second wafer to obtain the second notch image of the second notch; S140: Based on the second notch image, obtain the geometric feature points of the second notch as the first reference point of the second wafer; S150: The second original image acquisition device acquires a local arc original image of the second wafer containing the second arc edge, so as to obtain the second reference point of the second wafer and determine the reference vector of the second wafer, wherein the relative positional relationship between the first reference point and the second reference point of the first wafer is the same as the relative positional relationship between the first reference point and the second reference point of the second wafer. S160: By driving the first stage and the second stage to move relative to each other, the second wafer reference vector is aligned with the first wafer reference vector, so as to align the second wafer with the first wafer.

[0068] The wafer alignment method provided in this invention is based on the rigid body geometry two-point alignment principle. It aligns the second wafer with the first wafer by introducing reference vectors for the first and second wafers, thus providing a template-free adaptive alignment scheme that eliminates the need for template creation, storage, and iterative updates. Furthermore, the wafer alignment method provided in this invention abandons the easily disturbed Notch vertices and wafer center fitting, improving alignment accuracy, stability, and anti-interference capability.

[0069] Figure 6 This is a schematic flowchart of a wafer alignment method provided in other embodiments of the present invention. In some embodiments of the present invention, such as... Figure 6 As shown, the wafer alignment method is applied to a wafer bonding device and further includes the following steps: Before step S110: obtaining the geometric feature points of the first notch as the first reference point of the first wafer based on the first notch image, step S170: driving the first stage and the second stage to move relative to each other according to the tilt angle determined by the first notch image, so that the first notch is aligned with the standard direction; Before step S140: obtaining the geometric feature points of the second notch as the first reference point of the second wafer based on the second notch image, step S180: driving the first stage and the second stage to move relative to each other according to the tilt angle determined by the second notch image, so that the second notch is aligned with the standard direction; Step S110: Based on the first notch image, obtaining the geometric feature points of the first notch as the first reference point of the first wafer includes: S1101: After aligning the first notch with the standard direction, based on the first notch image, obtain the geometric feature points of the first notch as the first reference point of the first wafer; Step S140: Based on the second notch image, obtain the geometric feature points of the second notch as the first reference point of the second wafer, including: S1401: After aligning the second notch with the standard direction, based on the second notch image, obtain the geometric feature points of the second notch as the first reference point of the second wafer.

[0070] In the above embodiments of the present invention, before determining the first reference point of the wafer reference vector, the wafer is driven according to the tilt angle so that the notch is aligned with the standard direction, thereby avoiding the problem of asymmetric notch contour extraction caused by wafer tilt and improving the accuracy of wafer alignment.

[0071] Specifically, in combination Figure 4 In addition to the wafer bonding equipment in the present invention, some embodiments of the present invention provide a wafer alignment method. Figure 7 This is a schematic flowchart of a wafer alignment method provided in other embodiments of the present invention. Figure 8 yes Figure 7 The illustrated wafer alignment method is applied in a wafer bonding apparatus. In some embodiments of the present invention, such as... Figure 7 As shown, the wafer alignment method includes the following steps: S1000: Place the first wafer 50 on the second stage 20, such as... Figure 8 As shown in state A; S1100: The first raw image acquisition unit 30 acquires a raw image of the first wafer 50 held on the second stage 20 to obtain a first notch image of the first notch 51, such as... Figure 8 As shown in state B; S1200: Based on the tilt angle θ1 determined by the first notch image, drive the second stage 20 holding the first wafer 50 so that the first notch 51 is aligned with a standard direction, such as... Figure 8 As shown in states C and D; S1300: With the first notch 51 aligned with the standard direction, based on the first notch image, the geometric feature points of the first notch 51 are obtained as the first reference point P1 of the first wafer 50, such as... Figure 8 As shown in state E; S1400: For the first wafer 50 placed on the second stage 20, the second raw image acquisition unit 40 acquires a partial arc raw image of the first wafer 50 including the first arc edge 52, to obtain the second reference point P2 of the first wafer 50 and determine the first wafer reference vector V1, as follows. Figure 8 As shown in states F and G; S1500: Transfer the first wafer 50 to the first stage 10, and place the second wafer 60 on the second stage 20, as follows: Figure 8 As shown in state H; S1600: The first raw image acquisition unit 30 acquires a raw image of the second wafer 60 held on the second stage 20 to obtain a second notch image of the second notch 61, such as... Figure 8 As shown in state I; S1700: Based on the tilt angle θ2 determined by the second notch image, drive the second stage 20 holding the second wafer 60 so that the second notch 61 is aligned with the standard direction, such as... Figure 8 As shown in states G and K; S1800: With the second notch 61 aligned with the standard direction, based on the second notch image, the geometric feature points of the second notch 61 are obtained as the first reference point P3 of the second wafer 60, such as... Figure 8 As shown in state L; S1900: For the second wafer 60 placed on the second stage 20, the second raw image acquisition unit 40 acquires a local arc raw image of the second wafer 60 including the second arc edge, to obtain the second reference point of the second wafer 60 and determine the second wafer reference vector V2 by P4, wherein the relative positional relationship between the first reference point P1 and the second reference point P2 of the first wafer 50 is the same as the relative positional relationship between the first reference point P3 and the second reference point P4 of the second wafer 60, such as... Figure 8 The states M and N are shown in the diagram; S2000: By driving the second stage 20, the second wafer reference vector V2 is aligned with the first wafer reference vector V1, so as to align the second wafer 60 on the second stage 20 with the first wafer 50 on the first stage 10, as follows. Figure 8 The states O and P are shown in the diagram.

[0072] Figure 9 This is a schematic flowchart of a wafer alignment method provided in other embodiments of the present invention. In some embodiments of the present invention, such as... Figure 9 As shown, step S10: Obtaining the first notch image of the first notch on the first wafer, including: S101: Acquire a first raw image of the first wafer acquired by the first raw image acquisition unit; S102: Perform adaptive image enhancement on the first original image and delineate the region of interest where the first gap is located to obtain the first processed image; S103: Perform adaptive threshold segmentation on the first processed image to extract the first gap image; Step S30: Obtain a second notch image of the second notch on the second wafer, including; S301: Acquire a second original image of the second wafer acquired by the first original image acquisition unit; S302: Perform adaptive image enhancement on the second original image and delineate the region of interest where the second gap is located to obtain the second processed image; S303: Perform adaptive threshold segmentation on the second processed image to extract the second gap image.

[0073] In the above embodiments of the present invention, the first original image can be an original image of the entire first wafer acquired by the first original image acquisition device, or it can be a partial image of the first wafer acquired by the first original image acquisition device, wherein the partial image includes a notch in the first wafer. Similarly, the second original image can be an original image of the entire second wafer acquired by the first original image acquisition device, or it can be a partial image of the second wafer acquired by the first original image acquisition device, wherein the partial image includes a notch in the second wafer. Generally, the image at the location of the notch in the original image is darker. The embodiments of the present invention do not limit the field of view of the original image acquisition device. Generally, the original image acquisition device used for wafer alignment in a wafer bonding device is a high-precision small-field-of-view camera, that is, the acquired image is an image of a local area of ​​the wafer.

[0074] It is understandable that after defining the region of interest, the region of interest can be extracted through methods such as image masking and image cropping to obtain the corresponding processed image.

[0075] In the above embodiments of the present invention, by defining the region of interest where the notch is located, only the effective detection area is retained, which can reduce invalid calculations and background interference. At the same time, by using adaptive threshold segmentation, the dark area of ​​the wafer notch can be stably extracted.

[0076] Figure 10 yes Figure 9 The diagram shows a flow chart of step S102 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 10 As shown, step S102: performing adaptive image enhancement on the first original image and defining the region of interest where the first gap is located to obtain the first processed image, including: S1021: Based on the first original image, enhance the contrast of the image edge of the first notch to obtain a first preprocessed image; S1022: Smooth the noise in the first preprocessed image to obtain the second preprocessed image; S1023: Delineate the region of interest where the first gap is located in the second preprocessed image to obtain the first processed image.

[0077] Figure 11 yes Figure 9 The diagram shows a flow chart of step S302 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 11 As shown, step S302: Perform image adaptive enhancement on the second original image and delineate the region of interest where the second gap is located to obtain the second processed image, including: S3021: Based on the second original image, enhance the contrast of the image edge of the second notch to obtain a third preprocessed image; S3022: Smooth the noise in the third preprocessed image to obtain the fourth preprocessed image; S3023: Delineate the region of interest where the second gap is located in the fourth preprocessed image to obtain the second processed image.

[0078] Since the image color at the location of the gap is relatively dark in the original image, enhancing the contrast of the image edges makes the contrast at the location of the gap higher, thus enabling quick and accurate location of the gap in the image.

[0079] For example, in step S1023, during the process of delineating the region of interest where the first gap is located in the second preprocessed image, the region of interest where the first gap is located in the second preprocessed image is delineated by blob localization analysis.

[0080] Similarly, for example, in step S3023, during the process of delineating the region of interest where the second gap is located in the fourth preprocessed image, the region of interest where the second gap is located in the fourth preprocessed image is delineated by blob localization analysis.

[0081] Figure 12 yes Figure 9 Another schematic diagram of step S102 in the wafer alignment method shown. In some embodiments of the present invention, such as... Figure 12 As shown, step S1021: Based on the first original image, enhance the contrast of the image edge of the first notch to obtain a first preprocessed image, including: S10211: Based on the first original image, maximize the contrast of the image edge of the first notch by power transformation and grayscale stretching to obtain the first preprocessed image; S1022: Smooth the noise in the first preprocessed image to obtain a second preprocessed image, including: S10221: Morphological closing operation is used to smooth the noise in the first preprocessed image to obtain the second preprocessed image.

[0082] Figure 13 yes Figure 9 Another schematic diagram of step S302 in the wafer alignment method shown. In some embodiments of the present invention, such as... Figure 13 As shown, step S3021: Based on the second original image, enhance the contrast of the image edge of the second notch to obtain a third preprocessed image, including: S30211: Based on the second original image, the third preprocessed image is obtained by maximizing the contrast of the image edge of the second notch through power transformation and grayscale stretching. S3022: Smooth the noise in the third preprocessed image to obtain a fourth preprocessed image, including: S30221: Morphological closing operation is used to smooth the noise in the third preprocessed image to obtain the fourth preprocessed image.

[0083] In the above embodiments of the present invention, the contrast of the image edge is maximized by power transformation and grayscale stretching, so that the contrast at the location of the gap is higher, and the location of the gap in the image can be located more quickly and accurately.

[0084] Figure 14 yes Figure 9 The diagram shows a flow chart of step S103 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 14 As shown, step S103: performing adaptive threshold segmentation on the first processed image to extract the first gap image includes: S1031: Adaptive standard deviation local threshold segmentation is used to extract the dark region of the first gap from the first processed image to obtain the fifth preprocessed image; S1032: Filter out the interference in the fifth preprocessed image to obtain the sixth preprocessed image; S1033: By area filtering, separate the complete first target region of the first gap from the sixth preprocessed image; S1034: Perform morphological opening and closing operations and region difference on the image of the first target region to obtain the first gap image.

[0085] For example, after applying adaptive standard deviation local threshold segmentation to the first processed image, the resulting fifth preprocessed image may include multiple connected components (i.e., multiple segmented regions); area filtering can be used to determine the connected component with the largest area, which may include the chamfered projection region of the gap location; by performing morphological opening and closing operations on the connected component and performing region difference on the operation results, the region of the first gap, i.e., the first target region, can be determined.

[0086] Specifically, by performing a closing operation on the connected regions obtained through area filtering, the union of the notch chamfer projection region and the semi-circular notch region can be obtained. Then, an opening operation is performed on the closing operation result to smooth and denoise it, yielding the opening operation result. By performing a region difference between the opening operation result and the previously filtered connected regions, that is, subtracting the connected regions obtained through area filtering from the connected regions corresponding to the opening operation result (equivalent to the complement of the intersection of the opening operation result and the connected regions obtained through area filtering), the region of the first notch can be obtained.

[0087] Figure 15 yes Figure 9 The diagram shows a flow chart of step S303 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 15 As shown, step S303: performing adaptive threshold segmentation on the second processed image to extract the second gap image includes: S3031: Adaptive standard deviation local threshold segmentation is used to extract the dark region of the second gap from the second processed image to obtain the seventh preprocessed image; S3032: Filter out the interference in the seventh preprocessed image to obtain the eighth preprocessed image; S3033: By area filtering, separate the complete second target region of the second gap from the eighth preprocessed image; S3034: Perform morphological opening and closing operations and region difference on the image of the second target region to obtain the second gap image.

[0088] For example, after applying adaptive standard deviation local threshold segmentation to the second processed image, the resulting seventh preprocessed image may include multiple connected components (i.e., multiple segmented regions); area filtering can be used to determine the connected component with the largest area, which may include the chamfered projection region of the gap location; by performing morphological opening and closing operations on the connected component and performing region difference on the operation results, the region of the second gap, i.e., the second target region, can be determined.

[0089] Similarly, by performing a closing operation on the connected regions obtained through area filtering, the union of the notch chamfer projection region and the semi-circular notch region can be obtained. Then, an opening operation is performed on the closing operation result to smooth and denoise the result, yielding the final result. By performing a region difference between the opening operation result and the previously filtered connected regions, that is, subtracting the connected regions obtained through area filtering from the connected regions corresponding to the opening operation result (equivalent to the complement of the intersection of the opening operation result and the connected regions obtained through area filtering), the region of the second notch can be obtained.

[0090] In the above embodiments of the present invention, during area screening, based on the ratio between the size of the wafer notch and the field of view of the image acquisition device, only areas that conform to this ratio are screened in the image.

[0091] In the above embodiments of the present invention, the complete missing target area is accurately separated by area screening and region difference, and the incomplete and excessively small interference areas in the image are eliminated.

[0092] In the above embodiments of the present invention, by performing multiple rounds of circular opening and closing morphological operations, the interference of small dust particles on the surface of the micro wafer and burrs on the edge of the wafer is filtered out, which greatly improves the accuracy of repeated positioning.

[0093] In some embodiments of the present invention, driving the wafer to align the notch with a standard orientation based on the tilt angle of the notch determined from the notch image includes: Based on the gap image, determine the two target endpoints of the gap; Calculate the tilt angle of the gap based on the two target endpoints; The wafer is driven to rotate according to the tilt angle so that the notch is aligned with the standard direction.

[0094] In the above embodiments of the present invention, the tilt angle of the notch is calculated based on two target endpoints, which simplifies the calculation process and improves the accuracy of wafer alignment.

[0095] Figure 16 yes Figure 3 The diagram shows a flow chart of step S203 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 16 As shown, step S203: Based on the first tilt angle of the first notch determined from the first notch image, drive the first wafer so that the first notch is aligned with a standard direction, including: S2031: Based on the first gap image, determine the two first target endpoints of the first gap; S2032: Calculate the first tilt angle of the first gap based on the two first target endpoints; S2033: Based on the first tilt angle, drive the first wafer to rotate so that the first notch is aligned with the standard direction.

[0096] Figure 17 yes Figure 3 The diagram shows a flow chart of step S403 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 17 As shown, step S403: Based on the second tilt angle of the second notch determined by the second notch image, drive the second wafer so that the second notch is aligned with the standard direction, including: S4031: Based on the second gap image, determine the two second target endpoints of the second gap; S4032: Calculate the second tilt angle of the second notch based on the two second target endpoints; S4033: Drive the second wafer to rotate according to the second tilt angle so that the second notch is aligned with the standard direction.

[0097] In some embodiments of the present invention, determining the two target endpoints of the notch based on the notch image includes: using Canny subpixel edge detection to extract the first contour lines on both sides of the notch in the notch image; expanding the notch in the notch image to obtain the second contour line of the expanded notch; and determining the intersection of the first contour line and the second contour line as the two target endpoints. The step of calculating the tilt angle of the gap based on the two target endpoints includes: calculating at least two parameters among the height difference between the two target endpoints, the width between the two target endpoints, and the distance between the two target endpoints; and calculating the tilt angle of the gap based on the at least two parameters.

[0098] In the above embodiments of the present invention, the tilt angle of the notch is calculated based on the two target endpoints determined by the first contour line and the second contour line, which can improve the accuracy of wafer alignment.

[0099] Figure 18 yes Figure 16 The diagram shows a flow chart of steps S2031 and S2032 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 18 As shown, in some embodiments of the present invention, step S2031: determining the two first target endpoints of the first gap based on the first gap image includes: S20311: Using Canny subpixel edge detection, extract the first contour lines on both sides of the first gap in the first gap image; S20312: Expand and enlarge the first gap in the first gap image to obtain the second contour line of the expanded first gap; S20313: Determine the intersection of the first contour line of the first gap and the second contour line of the expanded first gap as the two first target endpoints; Step S2032: Calculate the first tilt angle of the first notch based on the two first target endpoints, including: S20321: Based on the two first target endpoints, calculate at least two parameters from the height difference between the two first target endpoints, the width between the two first target endpoints, and the distance between the two first target endpoints; S20322: Calculate the first tilt angle of the first notch based on the at least two parameters.

[0100] Figure 19 yes Figure 17 The diagram shows a flow chart of steps S4031 and S4032 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 19 As shown, in some embodiments of the present invention, step S4031: determining the two second target endpoints of the second gap based on the second gap image includes: S40311: Using Canny subpixel edge detection, extract the first contour lines on both sides of the second notch in the second notch image; S40312: Expand and enlarge the second gap in the second gap image to obtain the second contour line of the expanded second gap; S40313: Determine the intersection of the first contour line of the second notch and the second contour line of the expanded second notch as the two second target endpoints; Step S4032: Calculate the second tilt angle of the second notch based on the two second target endpoints, including: S40321: Based on the two second target endpoints, calculate at least two parameters from the height difference between the two second target endpoints, the width between the two second target endpoints, and the distance between the two second target endpoints; S40322: Calculate the second tilt angle of the second notch based on the at least two parameters.

[0101] In some embodiments of the present invention, such as Figure 8As shown in state C, Canny subpixel edge detection is used to extract the first contour lines L1 on both sides of the first gap 51 in the first gap image; the first gap 51 in the first gap image is expanded to obtain the second contour line L2 of the expanded first gap 51; the intersection points Isc1 and Isc2 of the first contour line L1 of the first gap 51 and the second contour line L2 of the expanded first gap 51 are determined as the two first target endpoints. Similarly, two second target endpoints can be determined.

[0102] In some embodiments of the present invention, such as Figure 8 As shown in state C, based on the coordinate values ​​of the two first target endpoints Isc1 and Isc2 of the first wafer 50, at least two parameters are calculated: the height difference h1 between the two first target endpoints Isc1 and Isc2, the width w1 between the two first target endpoints Isc1 and Isc2, and the distance d1 between the two first target endpoints Isc1 and Isc2. The tangent angle tanθ1 of the first tilt angle θ1 can be obtained from the ratio of the height difference h1 and w1. The sine angle sinθ1 of the first tilt angle θ1 can be obtained from the ratio of the height difference h1 and the distance d1. The cosine angle cosθ1 of the first tilt angle θ1 can be obtained from the ratio of the width w and the distance d1. The first tilt angle θ1 of the first wafer 50 can be obtained from any one of tanθ1, sinθ1, and cosθ1, i.e., θ1 = arctanθ1, arcsinθ1, and arcosθ1.

[0103] Similarly, to obtain the second tilt angle θ2 of the second wafer 60, as follows: Figure 8 As shown in state J.

[0104] It should be noted that the junction between the outline of the wafer notch and the outline of the non-notch is arc-shaped, i.e., a rounded transition. Therefore, it is difficult to directly determine the endpoint at the junction. By expanding the domain, the outline of the notch obtained by expansion can be made to intersect with the outline of the adjacent non-notch, thus making it easier to determine the target endpoint.

[0105] In some embodiments of the present invention, the initial tilt angle of the notch is calculated based on the at least two parameters; When the initial tilt angle is greater than 90°, the initial tilt angle is subtracted by 180° to obtain the tilt angle; When the initial tilt angle is less than -90°, the initial tilt angle is increased by 180° to obtain the tilt angle. When the initial tilt angle is greater than or equal to -90° and less than or equal to 90°, the initial tilt angle is taken as the tilt angle.

[0106] For example, the angle between the line segment determined by the endpoint values ​​on both sides of the notch and the horizontal axis (i.e., the tilt angle of the wafer notch) may be 1° or -179°. Through the above normalization correction, the tilt angle of the wafer notch can be uniformly constrained to [-90°, 90°], completely eliminating the problem of 0° or 180° reverse jump.

[0107] In the above embodiments of the present invention, the tilt angle of the wafer notch is uniformly constrained to [-90°, 90°], which completely eliminates the problem of 0° or 180° reverse jump and improves the accuracy of the wafer notch alignment with the standard direction.

[0108] Figure 20 yes Figure 18 The diagram shows a flow chart of step S20322 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 20 As shown, step S20322: calculating the first tilt angle of the first notch based on the at least two parameters includes: S203221: Calculate the first initial tilt angle of the first notch based on the at least two parameters; S203222: When the first initial tilt angle is greater than 90°, the first initial tilt angle is subtracted by 180° to obtain the first tilt angle; S203223: When the first initial tilt angle is less than -90°, the first initial tilt angle is added to 180° to obtain the first tilt angle; S203224: When the first initial tilt angle is greater than or equal to -90° and less than or equal to 90°, the first initial tilt angle shall be taken as the first tilt angle.

[0109] Figure 21 yes Figure 19 The diagram shows a flow chart of step S40322 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 21 As shown, step S40322: calculating the second tilt angle of the second notch based on the at least two parameters includes: S403221: Calculate the second initial tilt angle of the second notch based on the at least two parameters; S403222: When the second initial tilt angle is greater than 90°, the second initial tilt angle is subtracted by 180° to obtain the second tilt angle; S403223: When the second initial tilt angle is less than -90°, the second initial tilt angle is added to 180° to obtain the second tilt angle; S403224: When the second initial tilt angle is greater than or equal to -90° and less than or equal to 90°, the second initial tilt angle is taken as the second tilt angle.

[0110] In some embodiments of the present invention, the first reference point is the center of the largest inscribed circle of the gap, or the first reference point is the center of the circle obtained by transforming the outline of the gap through the vertex of the gap, or the first reference point is the midpoint of the line connecting the endpoints on both sides of the gap, or the first reference point is a position point at a distance of a first preset distance and a second preset distance from two specific straight lines of the gap image, respectively. The two specific straight lines are located on both sides of the vertex of the gap, and for each of the two specific straight lines, the specific straight line is the line that can cover the maximum number of target outline points on the outline of the gap on the corresponding side of the vertex of the gap.

[0111] In this embodiment of the invention, the first reference point can be any point that can uniquely determine the notch, and both the first wafer and the second wafer use the same type of first reference point. For example, the first reference point of the first wafer is the center of the largest inscribed circle of the first notch, and the first reference point of the second wafer is the center of the largest inscribed circle of the second notch; the first reference point of the first wafer is the center of the first circle obtained by morphological transformation of the outline of the first notch through the vertex of the first notch, and the first reference point of the second wafer is the center of the second circle obtained by morphological transformation of the outline of the second notch through the vertex of the first notch; or the first reference point of the first wafer is the midpoint of the line connecting the endpoints on both sides of the first notch, and the first reference point of the second wafer is the midpoint of the line connecting the endpoints on both sides of the second notch.

[0112] In some embodiments, the endpoints on both sides of the notch are not limited to sharp inflection points where the curvature of the contour changes abruptly. When the root of the notch smoothly transitions to the contours on both sides without obvious sharp corners, the endpoints on both sides are the transition boundary points between the notch side contour and the wafer contour. In a specific embodiment, contour points can be traversed from the vertex of the notch along both sides, and the contour tangent angle or curvature can be calculated point by point. A curvature threshold or tangent angle threshold can be preset. If the tangent direction of multiple consecutive contour points traversed after a certain point tends to be consistent with the direction of the wafer body contour, then that point can be regarded as the corresponding side endpoint.

[0113] For example, the vertex of the notch can be understood as the contour point that is most deeply recessed into the wafer. Taking the vertex of the notch as the dividing point, the contour points on the notch contour can be divided into two sets of contour points on either side. For each set of contour points on that side, multiple target contour points can be determined, and then the straight line that passes through the most target contour points can be determined as the specific straight line on that side. There exists a point in the image that satisfies the following condition: the distance between this point and one of the specific straight lines is a first preset distance, and the distance between this point and the other specific straight line is a second preset distance. The first preset distance and the second preset distance can be equal or unequal. It is understood that for the first notch image and the second notch image, each of the two specific straight lines corresponds one-to-one. For ease of description and understanding, the two specific lines in the first gap image are denoted as the first specific line and the second specific line, respectively. The first specific line is located on the first side of the gap (e.g., the left side), and the second specific line is located on the second side of the gap (e.g., the right side). The two specific lines in the second gap image are denoted as the third specific line and the fourth specific line, respectively. The third specific line is located on the first side of the gap, and the fourth specific line is located on the second side of the gap. If the distance between the first reference point of the first gap image and the first specific line is a first preset distance, and the distance between the first reference point and the second specific line is a first preset distance, then the distance between the first reference point of the second gap image and the third specific line is a first preset distance, and the distance between the first reference point and the fourth specific line is a second preset distance.

[0114] In the above embodiments of the present invention, the first reference point is unique, which lays a good foundation for the determination of the reference vector and can improve the accuracy of wafer alignment.

[0115] In this embodiment of the invention, when the first reference point is a position point at which the distance between it and two specific straight lines of the gap image is a first preset distance and a second preset distance, respectively, obtaining the geometric feature point of the gap as the first reference point based on the gap image includes: determining two target line segments in the gap image, each of the two target line segments being a partial outline of the gap in the gap image; for each of the two target line segments, determining the straight line that can pass through the most target outline points on the target line segment as the specific straight line corresponding to the target line segment, thereby determining the two specific straight lines in the gap image; and determining the point in the gap image at which the distance between the first specific straight line and the two specific straight lines is the first preset distance and the distance between the second specific straight line and the two specific straight lines is the second preset distance as the geometric feature point of the gap as the first reference point.

[0116] In some embodiments, a contour template can be used to match the gap image. The contour template is a pre-constructed matching template used to characterize the contour shape information of a standard gap. The contour template can be composed of local contour points of the standard gap, such as the shape of a local arc segment composed of partial contour points of the hypotenuse regions on both sides of the gap apex, without carrying the full contour information of the gap; or it can contain the complete contour shape information of the standard gap, while pre-marking the relative positions and contour segments of the two target line segments relative to the overall gap contour within the template, that is, marking the position range of the two target line segments on the standard gap contour on the template. The embodiments of the present invention do not specifically limit the type of contour template, and can flexibly select it according to the actual shape of the gap: when the contour features of the hypotenuses on both sides of the gap are clear and only two hypotenuse line segments need to be identified, a simplified template containing only local contour points of the gap can be selected to reduce the matching computation load; when the gap contour structure is complex and there are many interfering edges around the gap, a complete contour template carrying the complete gap contour and the position markings of the target line segments can be selected, thereby improving the accuracy of matching positioning and target line segment extraction. During matching, the contour template is translated, scaled, and rotated within the gap image for matching operations. The image region in the gap image with the highest shape similarity to the contour template is then found to obtain the matching result. Based on the positional mapping relationship between the contour template and the image in the matching result, two target line segments in the gap image can be obtained. These two target line segments are two contour line segments on the actual contour of the gap.

[0117] In other embodiments, the vertex of the gap can be obtained by traversing all contour points on the gap contour and filtering the position coordinates of each contour point. Using this vertex as a boundary, the gap contour can be divided into two sets of contour regions. For each side of the contour on both sides of the vertex, multiple candidate straight lines can be generated from all contour points on that side. The number of contour points covered by each candidate straight line on that side is counted, and the straight line that covers the most target contour points is selected as the specific straight line corresponding to that side. Thus, two specific straight lines are obtained on both sides of the gap vertex. In this case, the target contour points are all contour points on the corresponding side. In yet another embodiment, the user can manually draw two line segments on the gap contour of the gap image to mark the start and end points of the target line segment. In yet another embodiment, the curvature of each contour point of the gap contour in the gap image can also be calculated. When the curvature of a consecutive preset number of contour points is less than a preset curvature (the smaller the curvature, the closer to a straight line), the line segment formed by these contour points can be used as the target line segment. In some embodiments, after extracting the notch region, the minimum bounding rectangle of the notch region is solved; based on the center point of the minimum bounding rectangle and the minor axis, the measurement station positions and measurement angles of the measuring calipers corresponding to the two hypotenuses of the notch are determined; then, along the direction perpendicular to the minor axis, measurement sampling points are laid out on the left and right hypotenuses of the notch using the measuring calipers, and the laid-out sampling points are used as target contour points; a specific straight line can be obtained by performing line fitting on the target contour points. It can be understood that in this embodiment, the sampling area can be defined based on the measurement range of the measuring calipers, thereby determining the final target line segment.

[0118] It is understandable that for target contour points, either all contour points on the target line segment can be selected, or only a portion of the sampling points can be selected. Sampling points can be obtained by performing sub-pixel edge measurement using calipers, or by sub-pixel edge extraction, contour point selection, threshold boundary sampling, and other equivalent point-finding methods. A specific straight line can be fitted using the target contour points; this specific straight line is the one that passes through the maximum number of target contour points on the target line segment. "Passing through the target contour points on the target line segment" can be understood as the distance between the straight line and the target contour points being less than or equal to a preset distance threshold.

[0119] The above technical solution obtains a specific straight line representing the hypotenuse of the notch and solves for the first reference point based on the preset distance between the two specific straight lines. It does not rely on the tip of the notch and the stability of the obtained first reference point is high.

[0120] Figure 22 yes Figure 2The diagram illustrates step S201 of the wafer alignment method shown. In some embodiments of the present invention, the first reference point of the first wafer is a position point at which the distance between it and two specific straight lines of the first notch image is a first preset distance and a second preset distance, respectively; for example... Figure 22 As shown, step S201: Obtaining the geometric feature points of the first notch as the first reference point based on the first notch image includes: S2011: Determine two target line segments in the first gap image, each of the two target line segments being a partial outline of the first gap in the first gap image; S2012: For each of the two target line segments, determine the line that passes through the most target contour points on the target line segment as the specific line corresponding to the target line segment, so as to determine the two specific lines in the first gap image; S2013: The point in the first gap image that is at a distance of the first specific line from the two specific lines to the first preset distance and at a distance of the second specific line from the two specific lines to the second preset distance is determined as the geometric feature point of the first gap and is used as the first reference point.

[0121] Figure 23 yes Figure 2 The diagram illustrates step S401 of the wafer alignment method shown. In some embodiments of the present invention, the first reference point of the second wafer is a position point at which the distance between it and two specific straight lines of the second notch image is a first preset distance and a second preset distance, respectively; for example... Figure 23 As shown, step S401: Obtaining the geometric feature points of the second notch as the first reference point based on the second notch image includes: S4011: Determine two target line segments in the second gap image, each of the two target line segments being a partial outline of the second gap in the second gap image; S4012: For each of the two target line segments, determine the line that passes through the most target contour points on the target line segment as the specific line corresponding to the target line segment, so as to determine the two specific lines in the second gap image; S4013: The point in the second gap image that is at a distance of the first specific line from the two specific lines to the first preset distance and at a distance of the second specific line from the two specific lines to the second preset distance is determined as the geometric feature point of the second gap and is used as the first reference point.

[0122] In some embodiments of the present invention, the first reference point is the center of a circle obtained by morphological transformation of the outline of the gap, passing through the vertex of the gap. The step of obtaining the geometric feature points of the gap as the first reference point based on the gap image includes: The outline of the notch in the notch image is smoothed and denoised; The contour of the notch after smoothing and noise reduction is transformed to extract the circle passing through the vertex of the notch; Calculate the center of the circle and use the center as the first reference point.

[0123] In the above embodiments of the present invention, the center of the circle obtained by transforming the contour of the notch through the vertex of the notch is used as the first reference point. The first reference point is unique, which lays a good foundation for the determination of the reference vector. At the same time, by smoothing and reducing noise, the accuracy of subsequent processing of the notch contour can be improved, and the accuracy of wafer alignment can be further improved.

[0124] Figure 24 yes Figure 2 Another flowchart illustrating step S201 in the wafer alignment method shown. In some embodiments of the present invention, the first reference point of the first wafer is the center of a first circle obtained by morphological transformation of the contour of the first notch, passing through the vertex of the first notch; like Figure 24 As shown, step S201: Obtaining the geometric feature points of the first notch as the first reference point based on the first notch image includes: S2011': Smooth and reduce noise on the first contour of the first notch; S2012': Perform contour shape transformation on the first contour of the first gap after smoothing and noise reduction, and extract the first circle passing through the vertex of the first gap; S2013': Calculate the first center of the first circle and use the first center as the first reference point of the first wafer.

[0125] Figure 25 yes Figure 2 Another flowchart illustrating step S401 in the wafer alignment method shown. In some embodiments of the present invention, the first reference point of the first wafer is the center of the second circle obtained by morphological transformation of the contour of the second notch, passing through the vertex of the second notch; like Figure 25 As shown, step S401: Obtaining the geometric feature points of the second notch as the first reference point based on the second notch image includes: S4011': Smooth and reduce noise on the second contour of the second notch; S4012': Perform contour shape transformation on the second contour of the second gap after smoothing and noise reduction, and extract the second circle passing through the vertex of the second gap; S4013': Calculate the second center of the second circle and use the second center as the first reference point of the second wafer.

[0126] In some embodiments of the present invention, the second reference point is the intersection of the arc edge on one side of the wafer and the predetermined axis of the original image acquired by the second original image acquisition device.

[0127] The above embodiments of the present invention can quickly and accurately obtain the second reference point.

[0128] In this embodiment of the invention, the second reference point of the first wafer can be the intersection of the first arc edge of the first notch and the axis of the original image of the first wafer from the second original image acquisition device. Similarly, the second reference point of the second wafer is also the intersection of the second arc edge of the second notch and the axis of the original image of the second wafer from the second original image acquisition device. It should be noted that the relative positional relationship between the first and second reference points of the first wafer and the second reference point of the second wafer should be substantially consistent. Typically, during wafer loading, the relative positional deviation of the first and second wafers on the horizontal plane is small, and the misalignment of the arc edge included in the image acquired by the second original image acquisition device for the arc edge is small. Therefore, the relative positional deviation of the physical contour points on the corresponding wafers mapped from the intersection points of the arc edges of the two wafers with their corresponding axes on the horizontal plane can be ignored. Therefore, it can be considered that the relative position between the first and second reference points of the first wafer is consistent with the relative position between the first and second reference points of the second wafer.

[0129] This invention does not limit the specific method of contour shape transformation. In one specific embodiment, contour shape transformation may be performed using the function shape_trans.

[0130] In some embodiments of the present invention, obtaining the second reference point at a predetermined position on the arc edge of one side of the wafer includes: Acquire a local arc original image acquired by the second original image acquisition device, wherein the local arc original image includes the arc edge; Extract the arc edge from the original image of the local arc; The intersection of the arc edge and the predetermined axis is obtained as the second reference point, wherein the predetermined axis is the axis at a predetermined position in the width direction of the local arc original image in the camera coordinate system where the second original image acquisition device is located.

[0131] In the above embodiments of the present invention, the second reference point can be obtained quickly and accurately by acquiring the original image of the local arc.

[0132] Figure 26 yes Figure 2 The diagram shows a flow chart of step S202 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 24 As shown, step S202: Obtaining a second reference point at a predetermined position on the first arc edge of one side of the first wafer includes: S2021: Obtain a partial arc original image of the first wafer acquired by the second original image acquisition device, wherein the partial arc original image of the first wafer includes the first arc edge; S2022: Extract the first arc edge from the original image of the local arc of the first wafer; S2023: Obtain the intersection point of the first arc edge and the predetermined axis of the original partial arc image of the first wafer as the second reference point of the first wafer, wherein the predetermined axis of the original partial arc image of the first wafer is the axis at a predetermined position in the width direction of the original partial arc image of the first wafer in the camera coordinate system where the second original image acquisition device is located.

[0133] Figure 27 yes Figure 2 The diagram shows a flow chart of step S402 in the wafer alignment method. In some embodiments of the present invention, such as... Figure 25 As shown, step S402: Obtaining a second reference point at a predetermined position on the second arc edge of one side of the second wafer includes: S4021: Obtain a partial arc original image of the second wafer acquired by the second original image acquisition device, wherein the partial arc original image of the second wafer includes the second arc edge; S4022: Extract the second arc edge from the original image of the partial arc of the second wafer; S4023: Obtain the intersection point of the second arc edge and the predetermined axis of the original partial arc image of the second wafer as the second reference point of the second wafer, wherein the predetermined axis of the original partial arc image of the second wafer is the axis at a predetermined position in the width direction of the original partial arc image of the second wafer in the camera coordinate system where the second original image acquisition device is located.

[0134] In some embodiments of the present invention, the predetermined axis is the central axis along the width direction of the local circular arc original image in the camera coordinate system where the second original image acquisition device is located. For example, Figure 8 In states G and N, the predetermined axis is the central axis (i.e., the axis at the 1 / 2 position of the width direction) of the local arc original image in the camera coordinate system where the second original image acquisition device is located. The central axis of the width direction of the local arc original image is the clearest part of the image field of view; therefore, using this central axis can improve the accuracy of wafer alignment.

[0135] In other embodiments of the present invention, the predetermined axis is an axis located at 1 / 3 or 1 / 4 of the width direction of the third original image in the camera coordinate system where the second original image acquisition device is located.

[0136] Figure 28 This is a flowchart illustrating step S50 of the wafer alignment method provided in an embodiment of the present invention. In some embodiments of the present invention, such as... Figure 28 As shown, step S50: Aligning the second wafer reference vector with the first wafer reference vector to align the second wafer with the first wafer, includes: S501: Calculate the offset and angle deviation between the second wafer reference vector and the first wafer reference vector based on the second wafer reference vector and the first wafer reference vector; S502: Align the second wafer with the first wafer according to the offset and the angle deviation.

[0137] In the above embodiments of the present invention, by aligning the second wafer and the first wafer using the offset and angular deviation between the second wafer reference vector and the first wafer reference vector, the accuracy of wafer alignment can be improved.

[0138] Wherein, the first wafer reference vector is the reference vector in the coordinate system of the image acquired by the first original image acquisition device, and the second wafer reference vector is the reference vector in the coordinate system of the image acquired by the second original image acquisition device. Through coordinate transformation, the two coordinate systems can be unified into one coordinate system, and the coordinates of points P1-P4 in the same coordinate system can be calculated, such as P1(X1, Y1), P2(X2, Y2), P3(X3, Y3), and P4(X4, Y4). Based on P1 and P2, the first wafer reference vector V1(X2-X1, Y2-Y1) can be determined, and based on P3 and P4, the second wafer reference vector V2(X4-X3, Y4-Y3) can be determined. Furthermore, the offsets ΔX, ΔY, and angular deviation Δθ of the first wafer reference vector can be determined. , ΔX = X3 - X1, ΔY = Y3 - Y1.

[0139] Another aspect of the present invention provides a wafer bonding method. Figure 29 This is a flowchart illustrating a wafer bonding method provided in an embodiment of the present invention. Figure 29 As shown, the wafer bonding method includes: S300: Align the second wafer and the first wafer according to any one of the wafer alignment methods described above; S310: Bond the first wafer and the second wafer.

[0140] The wafer bonding method of this invention is based on the rigid body geometry two-point alignment principle. It determines the first wafer reference vector based on the first and second reference points of the first wafer, and determines the second wafer reference vector based on the first and second reference points of the second wafer. Then, it aligns the second wafer with the first wafer reference vector, providing an adaptive alignment scheme without template matching, eliminating template creation, storage, and iterative updates. Based on the rigid body geometry two-point alignment principle, it abandons the fitting of easily disturbed notch vertices and the center of the entire wafer, improving alignment accuracy and stability, and enhancing anti-interference capabilities.

[0141] Some specific embodiments of the present invention provide a wafer bonding method that requires no template training or storage. Through angle pre-correction, circular steady-state feature extraction, and fixed dual-point baselines, it achieves low-cost, high-stability, and high-precision wafer alignment. This method is based on a bonding platform comprising an upper wafer stage and a lower wafer stage, used to hold two wafers and two cameras, respectively.

[0142] Figure 30 This is a flowchart illustrating a wafer bonding method according to a specific embodiment of the present invention. Figure 30 The wafer alignment method includes the following steps: S1. Adaptive Image Enhancement and Region of Interest (ROI) Cropping The original wafer image is acquired, and edge contrast is maximized through power transformation and grayscale stretching. Morphological closing operations are used to smooth image noise. Blob localization analysis is used to delineate the local region of interest where the wafer notch is located, retaining only the effective detection area and reducing invalid operations and background interference.

[0143] S2, Adaptive Threshold Segmentation and Gap Region Purification Adaptive standard deviation local thresholding is used to obtain an image containing multiple connected components (i.e., multiple segmented regions). Area filtering is used to determine the connected component with the largest area, which includes the chamfered projection region of the notch location. Morphological opening and closing operations are performed on this connected component, and region differencing is performed on the operation results to determine the notch region. Specifically, adaptive standard deviation local thresholding is used to stably extract the dark region of the wafer notch; multiple rounds of circular opening and closing morphological operations are used to filter out interference from tiny dust particles and edge burrs; through region differencing and area filtering, the complete notch target region is accurately separated, and incomplete or excessively small interference regions are removed.

[0144] S3, Subpixel Edge Extraction and Initial Angle Coarse Calculation The purified Notch region is expanded; Canny subpixel edge detection is used to extract high-precision contour lines on both sides of the Notch; the intersection point of the high-precision contour line and the expanded Notch region is determined; based on the intersection point, the horizontal width of the contour and the height difference between the two endpoints are determined, and the initial tilt angle of the Notch is calculated. This is used to reflect the current wafer placement angle.

[0145] S4, Angle Normalization Correction The calculated initial angles are corrected using interval constraints: • When the angle is greater than 90°, subtract 180° from the angle; • When the angle is <-90°, add 180° to the angle; constrain all angles to the range of [-90°, 90°] to completely eliminate the problem of reverse jump between 0° and 180° in straight line angles; drive the platform to rotate slightly according to the corrected angle so that the Notch opening maintains a positive standard posture and eliminates the problem of asymmetric profile cutting caused by wafer tilt.

[0146] S5, Steady-state extraction of the center of the notch-compliant circle. After the notch posture correction is completed, the standard positive notch contour is smoothed and denoised. The conformal circle passing through the notch vertex is extracted by contour shape transformation. The radius range of the conformal circle is screened and abnormal fitting results are eliminated. The center of the area of ​​the conformal circle is calculated as the first stable reference point M1.

[0147] S6, Construction of dual-point baseline Using the camera coordinate system with the equipment fixedly installed as a reference, the image width axis is defined as a fixed reference axis; the feature of the arc edge (at 90° with the notch) on one side of the wafer is extracted, and the intersection of the arc edge and the fixed central axis of the image is calculated as the second fixed reference point M2; the global reference line of the wafer is uniquely determined by the two points M1 (the center of the compliant circle obtained by morphological transformation of the notch's contour) + M2 (the intersection of the arc edge), without the need to fit the center of the outer circle of the wafer throughout the process, thus avoiding the systematic error caused by the full circle fitting.

[0148] S7, Wafer Alignment and Lamination The offset and angular deviation of the two reference lines are calculated in real time. By adjusting the X and Y translation and θ rotation of the motion platform, the reference lines of the upper and lower wafers are completely overlapped, and micron-level high-precision alignment and bonding are achieved.

[0149] The method provided in the above embodiments of the present invention achieves: (1) Template-free adaptation: completely abandons matching methods such as shape templates and NCC grayscale templates, eliminating the need for template creation, storage, and iterative updates, adapting to batch production of wafers of the same specifications, and reducing equipment maintenance costs; (2) Strong feature stability: abandons the easily disturbed Notch vertices, and adopts the center of the compliant circle as the core feature point, resisting interference from edge breakage, dust, and uneven edge grayscale, and significantly improving the positioning repeatability accuracy; (3) Angle control closed loop: a unique ±90° interval angle normalization strategy is created to fundamentally solve the angle problem. (3) Jump problem, combined with the process of first correction and then fitting, to ensure the accuracy of the compliant circle extraction; (4) Simplified error source: the whole circle center fitting step of the wafer is eliminated, the first-level algorithm error is reduced, the algorithm process is simpler and the calculation speed is faster; (5) Mass production accuracy meets the standard: under normal workshop conditions, the alignment repeatability is stable at ≤20μm, which meets the requirements of 8-inch and 12-inch wafer bonding and advanced packaging process; (6) Rigorous and universal logic: based on the rigid body geometry double point line determination principle, the principle is self-consistent and without loopholes, and can be widely transferred to semiconductor equipment such as wafer thinning, cutting and testing.

[0150] As described above, another aspect of the present invention provides a wafer bonding apparatus, comprising: a first stage for holding a first wafer during wafer bonding, the first wafer having a first arc edge and a first notch for positioning; a second stage for holding a second wafer during wafer bonding, the second wafer having a second arc edge and a second notch for positioning; a first raw image acquisition unit for acquiring raw images of the first wafer held on the second stage and acquiring raw images of the second wafer held on the second stage; a second raw image acquisition unit for acquiring raw images of a portion of the first wafer held on the second stage including the first arc edge and acquiring raw images of a portion of the second wafer held on the second stage including the second arc edge; a processor; and a memory storing computer program instructions, wherein the processor reads and executes the computer program instructions to implement the method provided in any of the embodiments described above.

[0151] In some embodiments of the present invention, the processor may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement embodiments of the present invention.

[0152] The memory may include a large-capacity storage device for data or instructions. For example, and not limitingly, the memory may include a hard disk drive (HDD), a floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. In one instance, the memory may include removable or non-removable (or fixed) media, or the memory may be non-volatile solid-state storage. The memory may be internal or external to the integrated gateway disaster recovery device.

[0153] In some embodiments of the present invention, the memory may be a read-only memory (ROM). The ROM may be a masked read-only memory (MROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), an electrically erasable programmable read-only memory (EAROM), or flash memory, or a combination of two or more of these.

[0154] Another aspect of the present invention provides a computer-readable storage medium (such as a non-transitory computer-readable storage medium). The computer-readable storage medium stores computer program instructions; when executed by a processor, these computer program instructions implement the methods provided in any of the above embodiments.

[0155] Another aspect of the present invention provides a computer program product, including a computer program that, when processed for execution, implements the method provided in any of the above embodiments.

[0156] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0157] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, read-only memory (ROM), flash memory, erasable read-only memory (EROM), floppy disks, compact disc read-only memory (CD-ROM), optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0158] It should also be noted that the exemplary embodiments mentioned in this invention describe methods or systems based on a series of steps or apparatus. However, this invention is not limited to the order of the steps described above; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0159] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0160] The above description is merely a specific embodiment of the present invention. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the protection scope of the present invention.

Claims

1. A wafer alignment method, wherein the wafer has a notch for positioning, characterized in that, The method includes: Obtain the first notch image of the first notch on the first wafer; The first wafer reference vector of the first wafer is obtained by performing the wafer reference vector determination process; Obtain the second notch image of the second wafer; The second wafer reference vector of the second wafer is obtained by performing the wafer reference vector determination process. Align the second wafer reference vector with the first wafer reference vector to align the second wafer with the first wafer. The wafer reference vector determination process includes: The geometric feature points of the notch are obtained based on the notch image as the first reference point; A second reference point is obtained at a predetermined position on the arc edge of one side of the wafer, and the vector pointing from one of the first reference point and the second reference point to the other is determined as the wafer reference vector.

2. The wafer alignment method according to claim 1, characterized in that, Before obtaining the geometric feature points of the notch based on the notch image as the first reference point, the wafer reference vector determination process further includes: Based on the tilt angle of the notch determined from the notch image, the wafer is driven so that the notch is aligned with a standard orientation; The step of obtaining the geometric feature points of the notch as the first reference point based on the notch image includes: after driving the wafer to align the notch with the standard direction, obtaining a new notch image, and obtaining the geometric feature points of the notch aligned with the standard direction in the new notch image as the first reference point.

3. The wafer alignment method according to claim 1, characterized in that, The process of obtaining the first notch image of the first notch on the first wafer includes: Acquire a first raw image of the first wafer acquired by a first raw image acquisition device; The first original image is subjected to adaptive image enhancement, and the region of interest where the first gap is located is defined to obtain the first processed image; Adaptive threshold segmentation is performed on the first processed image to extract the first gap image; The process of obtaining the second notch image of the second wafer includes: Acquire a second raw image of the second wafer acquired by the first raw image acquisition device; The second original image is subjected to adaptive image enhancement, and the region of interest where the second gap is located is defined to obtain the second processed image; Adaptive thresholding is performed on the second processed image to extract the second gap image.

4. The wafer alignment method according to claim 3, characterized in that, The step of performing adaptive image enhancement on the first original image and defining the region of interest where the first gap is located to obtain the first processed image includes: Based on the first original image, the contrast of the image edge of the first notch is enhanced to obtain the first preprocessed image; Smooth the noise in the first preprocessed image to obtain the second preprocessed image; In the second preprocessed image, the region of interest where the first gap is located is defined to obtain the first processed image; The step of performing adaptive image enhancement on the second original image and defining the region of interest where the second gap is located to obtain the second processed image includes: Based on the second original image, the contrast of the image edges of the second notch is enhanced to obtain a third preprocessed image; The noise in the third preprocessed image is smoothed to obtain the fourth preprocessed image; The region of interest where the second gap is located is defined in the fourth preprocessed image to obtain the second processed image.

5. The wafer alignment method according to claim 4, characterized in that, The step of enhancing the contrast of the image edge of the first gap based on the first original image to obtain the first preprocessed image includes: maximizing the enhancement of the contrast of the image edge of the first gap by power transformation and grayscale stretching based on the first original image to obtain the first preprocessed image. The step of smoothing the noise in the first preprocessed image to obtain the second preprocessed image includes: smoothing the noise in the first preprocessed image using morphological closing operations to obtain the second preprocessed image; The step of enhancing the contrast of the image edge of the second notch based on the second original image to obtain a third preprocessed image includes: maximizing the enhancement of the contrast of the image edge of the second notch by power transformation and grayscale stretching based on the second original image to obtain the third preprocessed image; Smoothing the noise in the third preprocessed image to obtain the fourth preprocessed image includes: smoothing the noise in the third preprocessed image using morphological closing operations to obtain the fourth preprocessed image.

6. The wafer alignment method according to claim 3, characterized in that, The step of performing adaptive threshold segmentation on the first processed image to extract the first gap image includes: Adaptive standard deviation local threshold segmentation is used to extract the dark region of the first gap from the first processed image to obtain the fifth preprocessed image; The interference in the fifth preprocessed image is filtered out to obtain the sixth preprocessed image; The first target region of the complete first gap is separated from the sixth preprocessed image by area filtering; The first gap image is obtained by performing morphological opening and closing operations and region differencing on the image of the first target region. The step of performing adaptive threshold segmentation on the second processed image to extract the second gap image includes: Adaptive standard deviation local thresholding is used to extract the dark region of the second gap from the second processed image to obtain the seventh preprocessed image; The interference in the seventh preprocessed image is filtered out to obtain the eighth preprocessed image; An image of the complete second target region of the second gap is separated from the eighth preprocessed image by area filtering; The second gap image is obtained by performing morphological opening and closing operations and region difference on the image of the second target region.

7. The wafer alignment method according to claim 2, characterized in that, The step of driving the wafer to align the notch with a standard orientation based on the tilt angle of the notch determined from the notch image includes: Based on the gap image, determine the two target endpoints of the gap; Calculate the tilt angle of the gap based on the two target endpoints; The wafer is driven to rotate according to the tilt angle so that the notch is aligned with the standard direction.

8. The wafer alignment method according to claim 7, characterized in that, The step of determining the two target endpoints of the gap based on the gap image includes: Canny subpixel edge detection is used to extract the first contour lines on both sides of the notch in the notch image; The gap in the gap image is expanded and enlarged to obtain the second contour line of the expanded gap; The intersection of the first contour line and the second contour line is determined as the two target endpoints; The step of calculating the tilt angle of the gap based on the two target endpoints includes: Based on the two target endpoints, calculate at least two of the following parameters: the height difference between the two target endpoints, the width between the two target endpoints, and the distance between the two target endpoints; The tilt angle of the notch is calculated based on the at least two parameters.

9. The wafer alignment method according to claim 8, characterized in that, The step of calculating the tilt angle of the notch based on the at least two parameters includes: Calculate the initial tilt angle of the notch based on the at least two parameters; When the initial tilt angle is greater than 90°, the initial tilt angle is subtracted by 180° to obtain the tilt angle; When the initial tilt angle is less than -90°, the initial tilt angle is increased by 180° to obtain the tilt angle. When the initial tilt angle is greater than or equal to -90° and less than or equal to 90°, the initial tilt angle is taken as the tilt angle.

10. The wafer alignment method according to any one of claims 1-9, characterized in that, The first reference point is the center of the largest inscribed circle of the notch. Alternatively, the first reference point can be the center of a circle obtained by transforming the outline of the gap through the vertex of the gap. Alternatively, the first reference point can be the midpoint of the line connecting the endpoints on both sides of the notch. Alternatively, the first reference point is a position point at which the distance between the two specific lines of the gap image is a first preset distance and a second preset distance, respectively. The two specific lines are located on both sides of the vertex of the gap, and for each of the two specific lines, the specific line is the line that can pass through the target contour points on the contour of that side of the gap from the corresponding side of the vertex of the gap.

11. The wafer alignment method according to claim 10, characterized in that, The first reference point is a position point located at a distance of a first preset distance and a second preset distance from two specific straight lines in the notch image, respectively. The step of obtaining the geometric feature points of the gap as the first reference point based on the gap image includes: Two target line segments are identified in the gap image, each of the two target line segments being a partial outline of the gap in the gap image; For each of the two target line segments, the straight line that passes through the most target contour points on that target line segment is determined as the specific straight line corresponding to that target line segment, so as to determine the two specific straight lines in the gap image; The point in the gap image that is at a distance of the first specific line from the two specific lines and at a distance of the second specific line from the two specific lines is determined as the geometric feature point of the gap and is used as the first reference point.

12. The wafer alignment method according to claim 10, characterized in that, The first reference point is the center of a circle obtained by transforming the outline of the gap through the vertex of the gap. The step of obtaining the geometric feature points of the gap as the first reference point based on the gap image includes: The outline of the notch in the notch image is smoothed and denoised; The contour of the notch after smoothing and noise reduction is transformed to extract the circle passing through the vertex of the notch; Calculate the center of the circle and use the center as the first reference point.

13. The wafer alignment method according to any one of claims 1-9, characterized in that, The second reference point is the intersection of the arc edge on one side of the wafer and the predetermined axis of the original image acquired by the second original image acquisition device.

14. The wafer alignment method according to claim 13, characterized in that, The step of obtaining the second reference point at a predetermined position on one side of the arc edge of the wafer includes: Acquire a local arc original image acquired by the second original image acquisition device, wherein the local arc original image includes the arc edge; Extract the arc edge from the original image of the local arc; The intersection of the arc edge and the predetermined axis is obtained as the second reference point, wherein the predetermined axis is the axis at a predetermined position in the width direction of the local arc original image in the camera coordinate system where the second original image acquisition device is located.

15. The wafer alignment method according to claim 14, characterized in that, The predetermined axis is the central axis along the width direction of the local circular arc original image in the camera coordinate system where the second original image acquisition device is located.

16. The wafer alignment method according to any one of claims 1-9, characterized in that, The step of aligning the second wafer reference vector with the first wafer reference vector to align the second wafer with the first wafer includes: Based on the second wafer reference vector and the first wafer reference vector, calculate the offset and angle deviation between the second wafer reference vector and the first wafer reference vector; The second wafer is aligned with the first wafer based on the offset and the angle deviation.

17. The wafer alignment method according to any one of claims 1-9, characterized in that, The wafer alignment method is applied to a wafer bonding apparatus, the wafer bonding apparatus comprising: A first stage is used to hold one of the first wafer and the second wafer during wafer bonding, wherein the first wafer has a first arc edge and a first notch for positioning, and the second wafer has a second arc edge and a second notch for positioning; A second stage is used to hold one of the first wafer and the second wafer during wafer bonding; A first raw image acquisition device is used to acquire raw images of the first wafer and raw images of the second wafer; The second raw image acquisition device is used to acquire a local arc raw image of the first wafer containing the first arc edge and to acquire a local arc raw image of the second wafer containing the second arc edge. The wafer alignment method includes: The first raw image of the first wafer is acquired by the first raw image acquisition device to obtain the first notch image of the first notch; Based on the first notch image, the geometric feature points of the first notch are obtained as the first reference points of the first wafer; The second original image acquisition device acquires a local arc original image of the first wafer containing the first arc edge, so as to obtain the second reference point of the first wafer and determine the reference vector of the first wafer; The first original image acquisition device acquires an original image of the second wafer to obtain a second notch image of the second notch; Based on the second notch image, the geometric feature points of the second notch are obtained as the first reference point of the second wafer; The second original image acquisition device acquires a local arc original image of the second wafer containing the second arc edge, so as to obtain the second reference point of the second wafer and determine the reference vector of the second wafer; By driving the first stage and the second stage to move relative to each other, the second wafer reference vector is aligned with the first wafer reference vector, thereby aligning the second wafer with the first wafer.

18. A wafer bonding method, characterized in that, include: The wafer alignment method according to any one of claims 1-17 aligns the second wafer and the first wafer; The first wafer and the second wafer are bonded together.

19. A wafer bonding apparatus, characterized in that, The device includes: A first stage is used to hold one of a first wafer and a second wafer during wafer bonding, wherein the first wafer has a first arc edge and a first notch for positioning, and the second wafer has a second arc edge and a second notch for positioning; A second stage is used to hold one of the first wafer and the second wafer during wafer bonding; A first raw image acquisition device is used to acquire raw images of the first wafer and raw images of the second wafer; The second raw image acquisition device is used to acquire a local arc raw image of the first wafer containing the first arc edge and to acquire a local arc raw image of the second wafer containing the second arc edge. A processor and a memory storing computer program instructions, wherein the processor reads and executes the computer program instructions to implement the wafer alignment method as described in any one of claims 1-17.

20. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions that, when executed by a processor, implement the wafer alignment method as described in any one of claims 1-17.

21. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the wafer alignment method as described in any one of claims 1-17.