A wafer crack detection device and its clamping assembly

Abstract

This invention provides a wafer crack detection device and its clamping assembly. The clamping assembly includes a plate and a support portion extending through the plate. The support portion includes a first support port, a second support port, and a support surface extending from the first support port to the second support port, wherein the diameter of the first support port is larger than the diameter of the second support port. When this assembly is used for transmission imaging, the support surface acts as an integrated aperture that fits tightly against the outer periphery of the wafer, physically and without blind spots blocking stray light from the illumination source that propagates around the wafer edge. Through this integrated structural design, this invention fundamentally solves the serious edge light leakage and image overexposure problems caused by the use of gripper-type fixation in the prior art, making it possible to accurately detect defects in the wafer edge region with high signal-to-noise ratio and high contrast.

Description

Technical Field

[0001] This utility model relates to the field of semiconductor manufacturing, and more specifically, to a wafer crack detection device and its clamping assembly. Background Technology

[0002] In the complex manufacturing process of semiconductor integrated circuits, the integrity of the internal structure of the wafer, as the basic substrate material, is crucial. During the series of mechanical processing and heat treatment processes such as cutting, grinding, thinning, and rapid thermal annealing, tiny cracks, or "cracks," are easily generated inside the wafer due to stress concentration, thermal shock, or mechanical damage.

[0003] In transmissive imaging crack detection of wafers, stable and non-destructive wafer fixation is crucial for ensuring image quality. Currently, traditional wafer fixation methods mostly employ mechanical gripper structures, using multiple grippers distributed around the wafer's perimeter to hold it at the edge. However, this method suffers from severe "edge overexposure" in high-power transmissive imaging scenarios.

[0004] Specifically, the physical gap between the grippers provides a path for stray light from the illumination source to bypass the wafer edge and directly enter the imaging module. This results in bright rings with saturation and loss of detail appearing in the wafer edge area during imaging, severely reducing the signal-to-noise ratio and contrast of defect detection in this region. The wafer edge is precisely a region of stress concentration and high incidence of damage. The detection blind zone caused by the traditional gripper-type fixing method has become a key bottleneck restricting the accuracy and reliability of transmission crack detection. Utility Model Content

[0005] This utility model provides a wafer crack detection clamping assembly, which includes a plate and a support portion disposed through the plate. The support portion includes a first support port, a second support port, and a support surface extending from the first support port to the second support port, wherein the diameter of the first support port is larger than the diameter of the second support port.

[0006] Furthermore, this utility model also provides a wafer crack detection device, which includes: a plate; a support portion disposed through the plate, the support portion including a first support port, a second support port and a support surface extending from the first support port to the second support port, wherein the diameter of the first support port is larger than the diameter of the second support port; a laser light source configured to face the support portion; and an image acquisition device configured to face the support portion.

[0007] Compared with existing technologies, this invention fundamentally eliminates edge light leakage and overexposure problems through the aforementioned structural improvements, achieving blind-spot-free detection of the entire wafer surface (including the outermost edge). This is thanks to the wafer clamping assembly with a ring-shaped, continuous bearing surface structure. Its working mechanism is that the inner wall of the bearing portion of this assembly is no longer discrete contact points, but a complete, 360° uninterrupted solid inclined surface. When the wafer is placed on it, its entire outer periphery is tightly fitted to this inclined surface made of opaque material. This structure is functionally equivalent to a highly efficient optical aperture perfectly integrated with the wafer. Any stray light attempting to diffract or scatter from the light source below to the wafer edge is physically blocked and absorbed by this solid, opaque inclined surface, preventing it from forming a propagation path around the wafer. This ensures that only light that truly penetrates the wafer entity (including its edge areas) can enter the camera. Therefore, this design completely eliminates the root cause of edge overexposure artifacts, allowing the real defect information in the wafer edge area to be clearly imaged under high contrast, providing a reliable data foundation for accurate identification by subsequent algorithms. Attached Figure Description

[0008] Figure 1 This is a three-dimensional schematic diagram of the overall structure of an embodiment of the wafer crack detection device described in this utility model.

[0009] Figure 2 yes Figure 1 A front view of the device shown.

[0010] Figure 3 This is a three-dimensional structural diagram of an embodiment of the wafer clamping assembly, a core component of this utility model.

[0011] Figure 4 It is along Figure 3 A schematic diagram of the cross-sectional structure of the wafer clamping assembly plate cut by the AA line, which also shows the state in which the wafer is supported.

[0012] Figure 5 This is a schematic diagram of the internal structure of an embodiment of the lighting module in the device of this utility model.

[0013] Figure 6 This is a schematic diagram of the imaging optical path and effect of this utility model.

[0014] Figure 7 This is a cross-sectional structural schematic diagram of another embodiment of the wafer clamping assembly of this utility model, which integrates vacuum adsorption function.

[0015] List of reference numerals 1—Wafer; 100—Illumination module; 110—Laser source; 120—Collimating lens system; 130—Powell lens; 200—Wafer clamping and transport assembly; 210—Wafer clamping assembly; 211—Bearing surface; 212—Board; 2122—First surface; 2124—Second surface; 213—Bearing part; 2132—First bearing port; 2134—Second bearing port; 214—Vacuum pipe; 215—Vacuum adsorption hole; 220—Linear transport stage; 221—Fixed base; 222—Moving slide; 300—Imaging module; 310—Image acquisition unit; 320—Imaging lens system; 400—Illumination light; 410—Effective signal light. Detailed Implementation

[0016] The technical solution of this utility model will be clearly and completely described below with reference to the accompanying drawings.

[0017] Reference Figure 1 and Figure 2 This invention demonstrates the overall system architecture of a wafer crack detection device. The device mainly consists of three core subsystems: an illumination module 100 fixedly mounted at the lower part of the device frame, an imaging module 300 fixedly mounted at the upper part of the device frame, and a wafer clamping and transport assembly 200 located between the two for carrying and transporting the wafer 1 to be tested. Throughout the detection process, the spatial positions of the illumination module 100 and the imaging module 300 remain constant. The wafer clamping and transport assembly 200, through its internal linear transport stage 220, drives the wafer clamping assembly 210 carrying the wafer 1 between the illumination module 100 and the imaging module 300 along a preset, precise straight path (e.g., [missing information]). Figure 1 The linear transport platform 220 (shown in the X direction) moves at a constant velocity in a straight line, meaning it is equipped with a plate 212. The illumination module 100 emits a static, line-sweeping illumination beam upwards, as shown in... Figure 1 and Figure 2 As schematically shown through shaded areas and internal arrows, the length direction of the beam (e.g., the Y direction) spans the entire wafer diameter and is perpendicular to the direction of movement of wafer 1 (X direction), meaning the line-scan illumination beam propagates along the Z direction. The image acquisition unit 310 in the imaging module 300, located directly above wafer 1, uses a narrow linear sensor conjugate to the line-scan beam to capture the transmitted light image passing through wafer 1 in real time, line by line. As wafer 1 moves smoothly, the image acquisition unit 310 continuously acquires line images at an extremely high line scan frequency. These one-dimensional line images are then seamlessly stitched together according to the temporal relationship of the motion using an internal image processing unit or an external computer, ultimately reconstructing a complete, high-resolution two-dimensional transmission image of wafer 1.

[0018] The internal structure and working principle of the lighting module 100 will now be explained in more detail. (Refer to...) Figure 5 The lighting module 100 functions to provide the system with a uniform, stable, and extremely high energy density linear penetrating light source. It is not a simple strip LED light or tube, but a precisely designed optical system. In one embodiment, the core light-emitting element of the lighting module 100 is a laser source 110 positioned facing the support portion 213. Laser is chosen as the light source because of its high collimation, high monochromaticity, and high energy density, which are crucial for penetrating thick and rough-surfaced wafers. Specifically, a narrowband laser source with a center wavelength in the range of 1060 nm to 1080 nm can be selected. This wavelength falls within the near-infrared transmission window of semiconductor materials, has a relatively low absorption coefficient, and experiences less energy attenuation when penetrating semiconductor materials of the same thickness. The original beam emitted by the laser source 110 typically has a certain divergence angle and an irregular spot shape, such as in… Figure 5 The laser beam, schematically shown above and adjacent to the laser source 110, requires shaping via a collimating lens system 120 composed of multiple lenses. This collimating lens system 120 transforms the diverging laser beam into a nearly parallel Gaussian beam with a controllable beam diameter and a circular or elliptical cross-section, as shown in... Figure 5 The diagram is schematically shown above and adjacent to the collimating lens system 120.

[0019] Simply obtaining a collimated Gaussian beam is insufficient, as the energy distribution of a Gaussian beam is a bell-shaped curve, strong in the center and weak at the edges. If it is directly broadened into a line spot using a cylindrical mirror, the intensity distribution of this line spot will also be Gaussian, which will cause significant problems for subsequent image processing algorithms. For the same crack defect, when it sweeps across the center of the spot, the image grayscale value will be very low (dark) due to the high background light intensity; while when it sweeps across the edge of the spot, the image grayscale value will be relatively high (bright) due to the weak background light intensity. This makes setting a uniform and reliable defect determination threshold extremely difficult. To solve this problem, this embodiment introduces a key line spot generation optical element—a Powell lens 130. The collimated Gaussian beam is incident on the Powell lens 130. The Powell lens is a specially designed aspherical cylindrical lens with a complex, precisely calculated aspherical curvature profile at its apex. Unlike ordinary cylindrical lenses, Powell lenses, through their unique aspherical design, can precisely redistribute the energy of an incident light beam, "pushing" the energy from the center of the beam to both ends, as in... Figure 5The beam pattern, schematically shown above and adjacent to the Powell lens 130, forms a straight beam with a highly uniform intensity distribution and a "flat-top" profile on the target plane in the far field (i.e., the plane where the wafer is located). This uniformity of intensity ensures that the received illumination intensity is essentially consistent regardless of the defect's location on the scan line, resulting in stable and repeatable grayscale features, which greatly improves the accuracy and robustness of the algorithm detection.

[0020] Next, the core innovative component of this utility model—the wafer clamping assembly 210—will be described in detail. This wafer clamping assembly 210 ingeniously integrates the non-destructive mechanical clamping function of the wafer with the optical blocking function of stray light into a single structure. Firstly, regarding material selection, such as in… Figure 3 and Figure 4 As shown, the wafer clamping assembly 210's plate 212 can be manufactured from a single piece of polyetheretherketone (PEEK) through high-precision machining. PEEK, a high-performance specialty engineering plastic, is widely used in the semiconductor industry to manufacture wafer carriers and cassettes due to its several key properties. First, PEEK has a smooth, self-lubricating surface, making it less prone to generating microparticles when in contact with the hard texture of the wafer, thus meeting the stringent cleanliness requirements of semiconductor manufacturing. Second, PEEK exhibits strong corrosion resistance to various chemicals that may be encountered during the process and maintains dimensional stability over a wide temperature range, ensuring long-term reliability of clamping accuracy. Third, PEEK possesses sufficient rigidity to support the wafer under its own weight and acceleration without significant deformation, while its hardness is far lower than that of single-crystal silicon, ensuring that the wafer surface or edges will not be scratched under any contact conditions. Fourth, by selecting antistatic grade PEEK materials with added carbon fiber or other conductive fillers, the accumulation of static electricity caused by friction can be effectively avoided, preventing electrostatic discharge from damaging the precision circuit structure already formed on the wafer.

[0021] The core of this wafer clamping assembly 210 lies in its unique geometric configuration, namely the through-hole in the center of the plate 212, which serves as the support portion 213. For example... Figure 3 and Figure 4 As shown, the support portion 213 is not a simple straight-walled cylindrical hole. For a wafer 1 with a diameter of 300 mm, the upper opening diameter of the support portion 213 is designed to be greater than 300 mm (e.g., Figure 4 The lower opening diameter is designed to be less than 300 mm (e.g., 305 mm as shown in the figure), while the lower opening diameter is designed to be less than 300 mm (e.g., 305 mm as shown in the figure). Figure 4(295 mm shown). Thus, the inner wall of the support portion 213 naturally forms a continuous, annular, downwardly narrowing support surface 211. That is, the support portion 213 includes a first support opening 2132, a second support opening 2134, and a support surface 211 extending from the first support opening 2132 to the second support opening 2134, wherein the diameter of the first support opening 2132 is larger than the diameter of the second support opening 2134. Additionally, as shown in... Figure 3 and Figure 4 As shown, the first bearing port 2132 and the second bearing port 2134 can be circular. Additionally, as in... Figure 4 As shown, the plate 212 has a first surface 2122 and a second surface 2124 facing away from each other. The first bearing port 2132 is coplanar with the first surface 2122 of the plate 212, and the second bearing port 2134 is coplanar with the second surface 2124 of the plate 212. In this embodiment, the bearing surface 211 can have an inclination angle in the range of 30° to 60°. In a specific embodiment, the inclination angle α of the inclined surface is set to 45°, which is an angle that achieves a good balance between mechanical positioning stability and the convenience of robot arm handling. When the wafer 1 to be tested is placed on this assembly, its entire outer peripheral edge at the bottom rests uniformly and stably on this bearing surface 211, achieving reliable support for the wafer.

[0022] The key innovative value of this structure lies in the fact that it is not merely a mechanical support, but also a highly efficient, integrated optical aperture that is seamlessly integrated with the wafer. When using conventional grippers to hold wafer 1, the light source from the lower illumination module generates a large amount of non-collimated stray light. This stray light can easily bypass the wafer edge through the gaps between the grippers and directly enter the upper camera lens, causing severe overexposure and bright rings at the image edges, thus obscuring the effective signal light in that area. However, in this invention, as... Figure 6 As shown, the solid support surface 211, made of opaque PEEK material, fits tightly and seamlessly beneath the entire edge of the wafer 1. Any illumination light 400 from the illumination module 100, even... Figure 6 As not shown, non-vertical and scattered light generated within the system, attempting to bypass the wafer edge, is physically blocked and absorbed by the supporting surface 211 of this entity, preventing it from entering the imaging module 300 above. Thus, only the effective signal light 410 that truly penetrates the wafer 1 entity (including its outermost edge region) can be captured by the camera. In other words, the supporting surface 211, acting as an integrated aperture, physically blocks the illumination light 400 from propagating to the imaging module 300 along a path that does not pass through the wafer 1. This design physically and completely cuts off the leakage path of stray light, thereby eliminating edge overexposure artifacts and obtaining an image with clear edge details and true grayscale.

[0023] To further enhance performance and adapt to different application scenarios, this invention may also include various modified embodiments. For example, in one embodiment, the surface of the bearing surface 211 may undergo additional treatment. A micron-level rough structure can be formed on the surface of the bearing surface 211 by methods such as sandblasting, chemical etching, or laser micromachining; that is, the bearing surface 211 has a texture, making it an optically diffuse reflective surface. Alternatively, a certified, low-reflectivity black light-absorbing coating suitable for cleanroom environments can be coated onto the surface. Both methods aim to minimize specular reflection or directional scattering of stray light by the bearing surface 211 itself, further improving the ability to suppress stray light, which is a common and effective means of controlling stray light in optical system design.

[0024] When handling ultra-thin or warped wafers, simple gravity placement may not guarantee a perfect fit between the wafer edge and the support surface 211, potentially creating minute gaps that affect the sealing effect of the aperture. To address this issue, this invention proposes... Figure 7 Another embodiment is shown. In this embodiment, an annular vacuum channel 214 is machined inside the plate 212 of the wafer clamping assembly 210. Multiple tiny vacuum adsorption holes 215 are evenly distributed along the circumference on the bearing surface 211, and these vacuum adsorption holes 215 are connected to the internal vacuum channel 214. When an external vacuum pump evacuates air from the vacuum channel 214 through an interface (not shown), a negative pressure is generated at the vacuum adsorption holes 215. This suction force can actively and firmly adsorb the edge of the wafer 1 onto the bearing surface 211, ensuring a zero-gap fit between the two, thereby achieving an absolutely reliable optical seal under any circumstances.

[0025] Furthermore, the design of this utility model has good scalability. For example, by simply adjusting the upper and lower opening diameters of the support portion 213 and the overall external dimensions of the plate 212, wafer clamping assemblies suitable for wafers of different standard sizes (such as 200 mm, 150 mm, etc.) can be manufactured, while the core design concept and technical effect of using continuous inclined planes as integrated apertures are completely consistent.

[0026] As in Figure 1As shown, the wafer clamping assembly 210, which carries and holds wafer 1, is securely mounted on the movable slide 222 of the linear transport stage 220. The linear transport stage 220 provides smooth, precise, and controllable linear motion throughout the scanning imaging process. Its specific structure includes a fixed base 221 fixed to the equipment frame and a movable slide 222 that can move along a high-precision linear guide on the fixed base 221. To achieve ultimate speed stability, the drive system preferably uses a direct-drive linear motor, such as a coreless linear motor. This type of motor fundamentally eliminates backlash, elastic deformation, and cogging effects inherent in traditional lead screw or belt drives, ensuring pulsation-free uniform motion at both low and high speeds. To guarantee motion accuracy, the position feedback system uses a high-resolution linear optical encoder to measure the actual position of the movable slide 222 in real time and accurately, feeding the signal back to the motion controller to form a high-bandwidth closed-loop servo control, thereby ensuring long-term constant motion speed. The entire linear transport platform 220 is made of cleanroom-compatible materials (such as anodized aluminum and stainless steel) and can be equipped with dust covers or bellows covers to meet the stringent cleanliness requirements of semiconductor equipment.

[0027] Finally, as in Figure 1 As shown, the imaging module 300, located directly above wafer 1, is responsible for receiving transmitted light and generating digital images. It mainly includes an image acquisition unit 310 positioned facing the carrier 213 and a matching imaging lens system 320. The linear image sensor inside the image acquisition unit 310 is specially designed to have high quantum efficiency and responsivity for light, such as light with a wavelength of 1064 nm, used by the illumination module 100. The imaging lens system 320 is responsible for clearly and with low distortion imaging the object plane of wafer 1 onto the photosensitive surface of the linear array sensor. As mentioned earlier, one of the keys to the success of this device is the accuracy of spatiotemporal synchronization. Through a synchronization controller, the line scanning frequency of the image acquisition unit 310, i.e., the number of lines of linear images acquired per second, is strictly locked to the movement speed of the linear transport stage 220. The two must satisfy the following relationship: ,in It is the speed of the moving slide. It's the camera's line frequency, and This is the projected size of a single pixel of the camera onto the wafer plane (i.e., object-space resolution). By precisely satisfying this relationship, it can be ensured that for every row of images captured by the camera, the wafer moves forward by exactly one pixel's height. This precise spatiotemporal correspondence ensures that the final stitched 2D image is free from stretching, compression, overlap, and missing rows in the direction of motion, thus obtaining a high-quality image with accurate geometric dimensions and smooth grayscale transitions. This provides the most reliable raw data foundation for subsequent defect recognition algorithms.

[0028] In summary, the workflow of this invention can be summarized as follows: First, the wafer to be tested 1 is placed on the bearing surface 211 of the wafer clamping assembly 210 using automated equipment. Next, the device is activated, and the laser source 110 in the illumination module 100 emits light. After being shaped by the collimating lens system 120 and the Powell lens 130, a static laser line with uniform intensity spanning the wafer diameter is formed below the wafer. Subsequently, the linear transport stage 220 drives the wafer 1 at a preset constant speed. It begins to move perpendicular to the laser line direction. Simultaneously, the image acquisition unit 310 in the imaging module 300 moves in accordance with the speed... Precisely synchronized line frequency The system begins continuously acquiring one-dimensional line images. Throughout the scanning process, the continuous bevel structure of the wafer clamping assembly 210 acts as a highly efficient aperture, effectively blocking any stray light that might bypass the wafer edge. Once wafer 1 has completely scanned the static laser line, the image acquisition unit 310 stops acquiring images and stitches all the acquired line images into a complete, high signal-to-noise ratio, edge-free overexposure, and motion-distortion-free two-dimensional transmission image of the wafer in memory. Finally, this image is transmitted to the backend industrial computer, where image processing algorithms automatically identify, locate, classify, and measure internal defects such as cracks based on grayscale anomalies in the image.

[0029] It should be noted that the technical solutions described in this utility model can be combined arbitrarily without conflict.

[0030] The above description is merely a specific embodiment of this utility model, but the protection scope of this utility model is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this utility model should be included within the protection scope of this utility model. Therefore, the protection scope of this utility model should be determined by the protection scope of the claims.

Claims

1. A wafer crack detection clamping assembly, characterized in that, include: plate body; as well as A support portion is disposed through the plate body. The support portion includes a first support opening, a second support opening, and a support surface, wherein the support surface extends from the first support opening to the second support opening. The diameter of the first bearing port is larger than the diameter of the second bearing port.

2. The wafer crack detection clamping assembly according to claim 1, characterized in that, The plate is made of polyetheretherketone.

3. The wafer crack detection clamping assembly according to any one of claims 1 to 2, characterized in that, The bearing surface has an inclination angle in the range of 30° to 60°.

4. The wafer crack detection clamping assembly according to any one of claims 1 to 2, characterized in that, The bearing surface has a texture or is coated with a light-absorbing coating.

5. The wafer crack detection clamping assembly according to any one of claims 1 to 2, characterized in that, in: The plate body is equipped with a vacuum channel; and At least one vacuum adsorption hole is distributed on the bearing surface, and the vacuum adsorption hole is connected to the vacuum pipe.

6. The wafer crack detection clamping assembly according to any one of claims 1 to 2, characterized in that, The first and second bearing ports are circular.

7. The wafer crack detection clamping assembly according to any one of claims 1 to 2, characterized in that, The plate has a first side and a second side facing away from each other. The first bearing port is coplanar with the first side of the plate, and the second bearing port is coplanar with the second side of the plate.

8. A wafer crack detection device, characterized in that, include: plate body; A bearing portion is disposed through the plate body. The bearing portion includes a first bearing port, a second bearing port, and a bearing surface. The bearing surface extends from the first bearing port to the second bearing port, wherein the diameter of the first bearing port is larger than the diameter of the second bearing port. A laser light source is configured to face the support portion; and An image acquisition device is configured to face the support.

9. The wafer crack detection device according to claim 8, characterized in that, It also includes a linear transport platform, which is equipped with the plate and has a movement rate corresponding to the line scanning frequency of the image acquisition device.

10. The wafer crack detection apparatus according to any one of claims 8 to 9, characterized in that, The laser source has a center wavelength in the range of 1060 nm to 1080 nm.

Images