Wafer surface defect detection system and method based on high-speed line-scan camera

Abstract

This invention discloses a wafer surface defect detection system and method based on a high-speed line scan camera, belonging to the field of wafer defect detection technology. It includes: an illumination system for providing adjustable bright-field illumination and illumination angle, and forming adjustable-size line spots on the wafer surface through a line spot generation device; a wafer scanning assembly including a stage for supporting the wafer and a displacement control assembly for driving the stage movement; an imaging system including a bright-field imaging assembly; a high-speed line scan camera for high-speed scanning and imaging of the signals collected by the imaging system, and balancing the signal-to-noise ratio, dynamic range, and detection resolution through a programmable pixel merging function; and a control system for coordinating and controlling the various system components and achieving parameter optimization and online adaptive calibration. This invention can achieve high-resolution, high-sensitivity wafer surface micro / nano defect detection while ensuring high detection speed.

Description

Technical Field

[0001] This invention belongs to the field of wafer defect detection technology, specifically relating to a wafer surface defect detection system and method based on a high-speed line scan camera. Background Technology

[0002] With the continuous development of semiconductor manufacturing technology and the advancement of process nodes, large-size wafers are increasingly widely used in integrated circuit production. The detection of micro- and nano-scale defects on wafer surfaces plays a crucial role in ensuring chip yield, improving product quality, and reducing production costs. High-precision and high-efficiency surface defect detection technology has become a key link in the semiconductor manufacturing process.

[0003] Currently, optical imaging systems are widely used in the industry for wafer surface defect detection, mainly including various technical approaches such as bright-field detection and polarization detection. Traditional detection systems often use area array cameras with point or area light sources for illumination, achieving point-by-point or area-by-area detection of the wafer surface through mechanical or optical scanning. However, this detection method has significant bottlenecks in terms of imaging speed and resolution: on the one hand, the frame rate limitation of area array cameras makes it difficult to meet the high-efficiency requirements of online detection of large-size wafers; on the other hand, the uneven light intensity distribution and low energy utilization of traditional illumination methods make it difficult to achieve high-resolution imaging while ensuring detection speed.

[0004] In recent years, high-speed line scan cameras have gradually become important equipment in the wafer inspection field due to their advantages such as high resolution, high frame rate, and large dynamic range. Line scan cameras employ linear sensor arrays, enabling continuous high-speed scanning imaging. Theoretically, frame rates can reach tens of thousands of frames per second, significantly improving inspection efficiency. In particular, high-speed line scan cameras integrating TDI (Time Delay Integration) sensors, through multi-level integration technology, can significantly improve the signal-to-noise ratio and sensitivity of the imaging system in low-light and high-speed motion scenarios, providing a new technical approach for the detection of micro- and nano-defects on wafer surfaces.

[0005] However, in practical applications, detection systems based on high-speed line scan cameras still face numerous technical challenges. First, during high-speed scanning, the system is prone to image blurring and motion blur, primarily due to a mismatch between scanning speed and camera exposure time, as well as factors such as mechanical vibration and temperature drift. Second, the design of the illumination system directly affects the detection results; traditional illumination methods often suffer from uneven light intensity distribution, low energy utilization, and difficulty adapting to different surface structures. Furthermore, while TDI sensors can improve the signal-to-noise ratio, they may experience insufficient integration time and decreased charge transfer efficiency during high-speed scanning, affecting image quality.

[0006] The complexity of wafer surface structures further increases the difficulty of inspection. In modern semiconductor manufacturing, wafer surfaces contain both patterned regions (such as integrated circuit patterns and photoresist patterns) and unpatterned regions (such as bare silicon regions and metallized regions), and commonly feature various etching structures (such as deep trenches, shallow trenches, and vias). The reflection and scattering characteristics of different materials and structures vary significantly, placing higher demands on the illumination methods and sensitivity of imaging systems. For example, the complex geometry of patterned wafer surfaces generates multiple scattering and diffraction effects, while defect identification on unpatterned wafer surfaces mainly relies on differences in surface roughness and material properties.

[0007] To address the aforementioned issues, line spot illumination technology has emerged. Line spot illumination enables uniform illumination of the wafer surface, improving imaging contrast and making it particularly suitable for detecting minute defects. By shaping point or area light sources into linear spots, energy density and detection efficiency can be improved while ensuring illumination uniformity. Bright-field illumination helps identify changes in surface reflection or absorption characteristics and is suitable for detecting surface contamination, oxide layer defects, etc. How to flexibly switch illumination modes according to the characteristics of the object being inspected in a bright-field illumination system, thereby enhancing the detection capability for multiple types of defects, is a pressing issue in the current technological field.

[0008] It is worth noting that balancing detection speed and accuracy remains a core challenge in system design. In semiconductor manufacturing, wafer inspection typically needs to be completed within a limited timeframe to avoid disrupting the production line. Increasing inspection speed often leads to a decrease in imaging resolution or signal-to-noise ratio, thus affecting the ability to identify micro- and nano-defects; conversely, pursuing higher detection accuracy may reduce inspection efficiency, making it difficult to meet the real-time inspection requirements of large-size wafer production lines. This contradiction is particularly pronounced when inspecting micro- and nano-defects approaching the optical diffraction limit. According to the Rayleigh criterion, the theoretical resolution of an optical system is approximately 1.22λ / NA, where λ is the illumination wavelength and NA is the numerical aperture. In practical applications, due to factors such as system noise, mechanical vibration, and temperature drift, the actual achievable resolution is often lower than the theoretical value.

[0009] Furthermore, modern semiconductor manufacturing places higher demands on the stability and reliability of inspection systems. These systems need to maintain stable performance during long-term operation, including stability in illumination intensity, image quality, and positioning accuracy. Simultaneously, they must possess good environmental adaptability, capable of operating normally under adverse conditions such as temperature variations, vibration, and dust. These requirements present challenges to the system's mechanical, optical, electronic, and software algorithms.

[0010] Therefore, achieving high detection accuracy while maintaining high detection speed, and adapting to defect detection of various etched structures on patterned and unpatterned wafer surfaces, has become a key technical challenge that the semiconductor industry urgently needs to overcome. Developing a highly integrated, stable, and adaptable wafer surface defect detection system is of great significance for improving semiconductor manufacturing levels, reducing production costs, and ensuring product quality. Summary of the Invention

[0011] In view of the above, the present invention aims to solve the technical problems existing in the current large-size wafer surface defect detection process, such as the difficulty in balancing detection speed and detection accuracy, insufficient imaging signal-to-noise ratio, uneven illumination, and poor adaptability to complex surface structures. It provides a wafer surface defect detection system and method based on a high-speed line scan camera, which can achieve high-resolution and high-sensitivity micro-nano defect detection while ensuring high detection speed. It is suitable for online detection needs of patterned and unpatterned wafer surfaces and various etched structures.

[0012] To achieve the above-mentioned objectives, the present invention provides the following technical solution:

[0013] This invention provides a wafer surface defect detection system based on a high-speed line scan camera, comprising:

[0014] An illumination system for providing bright-field illumination with adjustable illumination angle, and for forming adjustable-size line spots on the wafer surface through a line spot generating device therein;

[0015] The wafer scanning assembly includes a stage for carrying the wafer and a displacement control assembly for driving the stage to move. The displacement control assembly achieves full-coverage scanning of the wafer surface through a position feedback function.

[0016] An imaging system, including a bright-field imaging component for collecting reflected signals under bright-field illumination conditions;

[0017] High-speed line scan camera is used to perform high-speed scanning imaging of signals collected by the imaging system, and balances the signal-to-noise ratio, dynamic range and detection resolution through programmable pixel merging function;

[0018] The control system coordinates and controls the lighting system, wafer scanning components, imaging system, and high-speed line scanning camera, and optimizes the parameters of the detection system and performs online adaptive calibration based on the image shift modulation transfer function.

[0019] Preferably, the illumination system includes an illumination source for providing an illumination beam, a waveplate for adjusting the polarization state, a first attenuator for adjusting the illumination power, a beam expander for adjusting the beam size and shape, and the line spot generating device arranged sequentially along the optical path, thereby achieving bright-field illumination of the wafer surface by controlling the incident angle of the illumination beam.

[0020] Preferably, the line spot generating device under bright field illumination conditions adopts a cylindrical lens group, a diffractive optical element or a combination of the two, or a combination of an independently designed Fourier transform lens and a slit aperture.

[0021] Preferably, the displacement control component integrates a high-precision closed-loop control position feedback device to support the execution of a rotating spiral or line-by-line serpentine scanning path.

[0022] Preferably, the bright-field imaging assembly includes a microscope objective, a tube endoscope, a filter, and a re-imaging assembly. The re-imaging assembly includes a first objective, an analyzer, a second attenuator, and a second objective. Furthermore, an aperture is provided after the filter in the bright-field imaging assembly to adjust the numerical aperture of the microscope objective.

[0023] Preferably, the programmable pixel merging function of the high-speed linear scanning camera includes merging parallel to the scanning direction, merging perpendicular to the scanning direction, or a combination of both, and the merging objects include charge domain pixels, digital domain pixels, or a combination of both.

[0024] Preferably, when performing pixel merging perpendicular to the scanning direction, the effective pixel width after merging is matched with the optical resolution of the imaging system. The matching relationship is that the product of the optical Airy disk radius of the imaging system and the magnification of the imaging system is approximately equal to the effective pixel width after merging.

[0025] Preferably, the parameter optimization of the detection system based on the image shift modulation transfer function includes:

[0026] A multivariate objective function, which includes at least the optical imaging link, motion blur, pixel sampling effect, TDI integral series, and motion synchronization error, is constructed as the image shift modulation transfer function, expressed as:

[0027] ,

[0028] in, Represents the image-shift modulation transfer function. The function represents the modulation transfer function of a purely optical imaging link. Indicates the width of the line spot. This indicates the speed at which the wafer surface moves relative to the line scan direction. This indicates the effective exposure time of a high-speed line scan camera. This indicates the pixel length along the scan direction. This represents the spatial frequency component perpendicular to the line scan direction. This indicates the vertical dimension of a single TDI pixel. This represents the integral series of TDI. This represents the equivalent width of a pixel in the scanning direction. This is used to describe the normalized velocity error of the displacement control component during the integral period, which is also known as the motion synchronization error. The function is used to characterize the frequency response attenuation introduced by finite aperture or finite sampling. By co-optimizing the above parameters, a parameter combination that meets the preset high speed and high precision requirements can be obtained.

[0029] Preferably, online adaptive calibration of the detection system based on the image shift modulation transfer function includes:

[0030] During online inspection, the image shift modulation transfer function (EMF) curve of the inspection system is measured in real time or at regular intervals by scanning a known structure on a standard resolution plate or wafer. When the response value of the EMF curve is lower than a preset threshold, the exposure time is automatically adjusted. Scanning speed Linear spot width or pixel equivalent width One or more parameters in the image are adjusted until the response value of the image shift modulation transfer function curve recovers to the target level, thereby achieving dynamic compensation of the detection system performance.

[0031] To achieve the above-mentioned objectives, this invention also provides a wafer surface defect detection method based on a high-speed line scan camera, implemented using the aforementioned wafer surface defect detection system based on a high-speed line scan camera, comprising the following steps:

[0032] S1, using an illumination system to generate a line spot of light to illuminate the wafer surface;

[0033] S2, using the wafer scanning component to drive the wafer to move relative to the line spot, to achieve full coverage scanning of the wafer surface;

[0034] S3 uses an imaging system to collect the reflected signals from the wafer surface during the scanning process and image them onto a high-speed line scan camera;

[0035] S4 controls the high-speed line scan camera, which configures the pixel merging mode and level in real time according to the characteristics of the defect to be detected and the detection speed requirements, and acquires and processes the signal to generate a wafer surface image.

[0036] S5 analyzes the generated images to identify defects and optimizes the parameters of the detection system and performs online adaptive calibration through the control system.

[0037] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0038] (1) By introducing a programmable pixel merging function, this invention enables high-speed line scanning cameras to flexibly configure pixel merging modes and levels at the hardware (charge domain) or software (digital domain) level, based on specific detection scenarios (such as detecting large-area macroscopic defects or micro / nano defects) and speed requirements. For example, when pursuing high detection throughput, charge domain pixel merging can be used to significantly improve signal sensitivity and signal-to-noise ratio, ensuring the effective capture of weak scattering signals; when high resolution is required to capture the morphology of fine defects, merging can be reduced or a non-merging mode can be used. This flexibility allows the detection system to achieve the best balance between speed and accuracy, effectively balancing high resolution and high sensitivity, and meeting the detection needs of different process nodes.

[0039] (2) This invention innovatively uses the image shift modulation transfer function as a unified evaluation standard and optimization basis for the system's imaging quality. This function integrates multiple factors such as optical transfer function, scanning motion blur, TDI integral series, pixel size, and displacement synchronization error. During the design phase, key system parameters can be optimized collaboratively based on this function to ensure that the system maintains sufficient contrast (e.g., IMMTF value not less than 0.35) near the Nyquist frequency, thus guaranteeing high-definition imaging. During the operation phase, by monitoring the image shift modulation transfer function curve in real time, the system can automatically identify performance degradation and dynamically adjust parameters for compensation, effectively offsetting system performance drift caused by factors such as equipment aging and environmental fluctuations, ensuring high precision and high stability during long-term operation, and reducing reliance on manual calibration.

[0040] (3) The power, polarization state, angle and spot size of the illumination system provided by the present invention can be adjusted independently, and the numerical aperture and filter of the imaging system can also be configured. This flexibility, combined with the programmable scanning path, can optimize the signal excitation and collection efficiency for different wafer materials, film layers and various defects, and significantly improve the defect detection rate and process adaptability of the system. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a schematic diagram of the optical path structure of a wafer surface defect detection system based on a high-speed line scan camera provided in an embodiment of the present invention;

[0043] Figure 2 This is a schematic diagram of the bright field illumination optical path provided in an embodiment of the present invention;

[0044] Figure 3 This is a schematic diagram of the displacement control component provided in an embodiment of the present invention;

[0045] Figure 4 This is a schematic diagram of a spiral wafer scanning method and a serpentine wafer scanning method implemented using a displacement control component, provided in an embodiment of the present invention;

[0046] Figure 5 This is a schematic diagram of how the line spot illuminates the wafer surface in a vertical manner, and after reflection, it is collected by a microscope objective and forms a light spot on the image plane of a high-speed line scanning camera.

[0047] Figure 6 This is a schematic diagram of the structure of a single merged pixel under three different pixel merging methods after pixel merging is performed by the high-speed line scan camera provided in the embodiment of the present invention.

[0048] Figure 7 This is a schematic flowchart of a wafer surface defect detection method based on a high-speed line scan camera provided in an embodiment of the present invention.

[0049] The specific symbols in the attached diagram are as follows: 1. Illumination source; 2. Vertical illumination device; 2-1. Waveplate; 2-2. First attenuator; 2-3. Beam expander; 2-4. Bright field line spot generating device; 3. Microscope objective; 4. Tube mirror; 5. Filter; 6. Aperture; 7. Beam splitter; 8. First objective; 9. Analyzer; 10. Second attenuator; 11. Second objective; 12. High-speed line scan camera; 13. Wafer stage; 14. Displacement control assembly; 14-1. X / Y direction displacement control assembly; 14-2. Wafer stage rotation control assembly; 15. Wafer under test. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the scope of protection of this invention.

[0051] The inventive concept of this invention is to address the problems existing in current large-size wafer surface defect detection, such as the difficulty in balancing detection speed and accuracy, insufficient imaging signal-to-noise ratio, uneven illumination, and poor adaptability to complex surface structures. This invention proposes a wafer surface defect detection system and method based on a high-speed line scan camera. This system achieves efficient and high-precision online detection of patterned and unpatterned wafer surfaces and various etched structures through multi-mode illumination, line spot adjustment, TDI sensor pixel merging, intelligent displacement control, and collaborative optimization of the control system.

[0052] Based on the above-described inventive concept, this embodiment provides a wafer surface defect detection system and method based on a high-speed line scan camera, which is used for rapid and high-precision detection of micro- and nano-defects on large-size wafers. The detected micro- and nano-defects specifically refer to defects whose size is close to or even smaller than the optical diffraction limit. The optical diffraction limit is typically defined according to the classical Airy disk theory, i.e., given the numerical aperture (NA) of the imaging system and the illumination wavelength (λ), the smallest resolvable feature size is approximately 1.22λ / NA. For the detection system involved in this invention, by optimizing the illumination wavelength, increasing the numerical aperture of the imaging system, and employing a high-sensitivity TDI line scan camera, effective detection and identification of micro- and nano-defects approaching or even below this diffraction limit size can be achieved, thereby meeting the practical needs of high-precision defect detection on large-size wafer surfaces.

[0053] The optical path structure of the entire detection system is as follows: Figure 1 As shown, the system includes, but is not limited to, the following components:

[0054] The illumination source 1 can be a laser, plasma source, fiber optic bundle source, or other suitable light source type, and the wavelength and power parameters can be flexibly selected according to the detection requirements. The vertical illumination device 2 is used to project the beam emitted by the illumination source 1 in a near-vertical manner (with a slight tilt angle) onto the brightfield line spot generating device 2-4, achieving brightfield illumination. The specific structure and parameters of the above illumination device can be adjusted according to the actual application scenario, and a combination of multiple illumination methods can also be used.

[0055] The microscope objective 3 is used to collect scattered or reflected light from the wafer surface and its defects. The wafer can be a patterned wafer, a patternless wafer, or contain various etched structures. The tube mirror 4 is used in conjunction with the microscope objective 3 to generate an intermediate image. The filter 5, which can be installed at the back focal plane of the microscope objective 3 or moved to other suitable positions via a relay mirror system, is used to filter out signals of specific frequencies or types to improve image quality and defect identification capabilities. The aperture 6 is used to adjust the effective numerical aperture of the microscope objective 3, enabling flexible control of imaging resolution and depth of field. The beam splitter 7 functions to separate or guide the beam at different ratios; the specific beam splitting method and ratio can be flexibly set according to system requirements.

[0056] The first objective lens 8, analyzer 9, second attenuator 10, and second objective lens 11 together form the re-imaging assembly, which can work in conjunction with the microscope objective lens 3 and the tube endoscope 4 to magnify, transfer, or adjust the intermediate image. The re-imaging assembly supports switching of imaging magnification, which can be continuously adjustable or discretely multi-level, and can be implemented mechanically, photoelectrically, or using other suitable methods. The analyzer 9 and the second attenuator 10 are used to filter specific polarization state light signals collected by the microscope objective lens 3 and to finely adjust the light intensity to adapt to the imaging requirements of different detection scenarios. The area within the dashed box represents the entire bright-field imaging assembly; its specific structure and layout can be flexibly configured according to the detection object and system integration requirements.

[0057] The high-speed line scan camera 12 is used for high-speed, high-sensitivity image acquisition of signals collected by the imaging system. The high-speed line scan camera 12 employs a TDI (Time Delay Integration) sensor, which can be a CMOS TDI, CCD TDI, or other suitable type. It also features multiple functions such as pixel binning, dynamic range extension, and noise suppression. Specific imaging parameters and operating modes can be flexibly configured according to detection needs to adapt to diverse detection scenarios with different resolution, sensitivity, and speed requirements.

[0058] The wafer stage 13 and the displacement control component 14 together constitute the wafer scanning component. Large-size wafers achieve efficient surface defect detection through line spot illumination. The wafer is securely fixed on the wafer stage, which can be a vacuum adsorption type, mechanical clamping type, electromagnetic adsorption type, or other suitable fixing method. Full-surface scanning of the wafer is completed collaboratively by the displacement control component and the control system, enabling multiple scanning paths and motion modes to adapt to wafers of different sizes, shapes, and detection requirements.

[0059] like Figure 2As shown, the bright-field illumination optical path includes, but is not limited to, the following components: Illumination source 1, which can be a laser source, plasma source, or fiber bundle source, with a wavelength range covering the deep ultraviolet to near-infrared band, and the output power can be adjusted according to detection requirements. Waveplate 2-1, which can be a polarization state generator composed of a combination of a half-wave plate and a quarter-wave plate, or a liquid crystal phase delayer to achieve programmable polarization control, is used to convert the incident light into specific linearly polarized light, circularly polarized light, or elliptically polarized light to enhance the response capability to scattering signals from specific orientation defects on the wafer surface. First attenuator 2-2, which can be a continuously adjustable neutral density filter or a light intensity feedback control system based on an electro-optic modulator (such as an acousto-optic modulator), is used to precisely adjust the light power irradiated onto the wafer surface, avoid camera target saturation, and ensure signal linearity. Its dynamic range is preferably adjustable with an optical density OD value of 0-4, and its stability is better than ±1%. The beam expander 2-3 can be a beam expanding system composed of a pair of aspherical lenses or a microlens array homogenizing device, used to shape the Gaussian distributed beam into a flat-top distribution, and to achieve continuous adjustment of the spot diameter within the range of 1 mm to 10 mm, with light intensity uniformity optimized to over 90%. The bright-field line spot generating device 2-4 focuses the incident light into a line spot through a cylindrical lens group, diffractive optical elements, or a combination thereof. Preferably, a double-linked aberration-correcting cylindrical lens structure is used to reduce optical distortion, so that the line spot forms a uniform illumination line on the wafer surface with an adjustable width of 1 μm–20 μm, a length covering the width of the camera target surface, and an edge steepness greater than 85%. The wafer under test 15 has the incident light in the bright-field illumination path tilted at a small angle of 1°–5° to the normal to the wafer surface to optimize the acquisition of reflection signals from micro-nano defects.

[0060] like Figure 3 As shown, the displacement control component 14 includes: an X / Y direction displacement control component 14-1, which can be a high-precision air-bearing platform driven by a linear servo motor, integrating a grating ruler or laser interferometer to form a fully closed-loop position feedback, with a positioning accuracy better than ±0.1μm and a repeatability better than ±0.05μm, capable of executing a serpentine scanning path; and a wafer stage rotation control component 14-2, which can be a direct-drive rotary motor in conjunction with a high-precision angle encoder (resolution not less than 0.1 arcseconds) to achieve continuous rotation of the wafer around the normal axis (speed range 0.1–10 rpm) to support helical scanning mode. Through the coordinated operation of the above components, the system can achieve high-speed, high-resolution bright-field detection of the wafer surface.

[0061] like Figure 4 As shown, the scanning methods include, but are not limited to, spiral scanning ( Figure 4 (a) and snake scan ( Figure 4(b) can also be used according to actual needs. The wafer scanning path includes the upper edge, center line, and lower edge, and the actual scanning range is determined by the length of the illumination line spot. The illumination line spot forms a line structure extending along the Y-axis on the wafer surface. The shape can be rectangular, elliptical, or other suitable geometric shapes to adapt to different types of lighting sources and system integration requirements. The distance between the line spot and the center of the wafer stage is R. During the detection process, the wafer rotates at an angular velocity ω and moves horizontally at a speed v. The speed v can be dynamically adjusted according to the change of R to optimize scanning coverage and detection efficiency. To ensure that no part of the wafer surface is missed during the scanning process, the effective scanning area is preferably set to 70% to 95% of the actual scanning area. The specific ratio can be flexibly set according to parameters such as wafer size, line spot width, and system resolution. For serpentine scanning paths, it is also necessary to ensure that the effective scanning area is smaller than the actual scanning area to avoid detection blind spots caused by lateral movement or path switching.

[0062] The scanning methods and parameters described above can be adjusted according to different application scenarios, wafer types, and detection accuracy requirements. Multiple scanning strategies can also be combined to achieve full coverage and high-precision automated detection of complex structures and large-size wafers. This invention is not limited to the specific embodiments described above; related scanning paths, motion parameters, and spot shapes can be flexibly configured according to actual needs to adapt to the diversified development of future processes and equipment.

[0063] In this embodiment, the line spot can be used to illuminate the wafer pattern surface perpendicularly. After scattering by the surface and defects, the signal is collected by the microscope objective 3 and finally formed on the target surface of the high-speed line scanning camera 12 to form a wafer image. The wafer image can be set according to system parameters to fill or even slightly exceed the width of the camera target surface, such as... Figure 5 As shown in (a), the target surface of the line scan camera, a single pixel of the line scan camera, and the image formed on the target surface by the illumination line spot after passing through the bright-field imaging component are illustrated. The width of the illumination line spot directly affects the surface width covered in each scan. For some applications, it is preferable to fill the width of the camera target surface with the wafer image to improve detection efficiency and coverage.

[0064] When the wafer image fills the width of the camera target surface, the scanning spacing w should match the length h of a single pixel to avoid image blurring and information loss caused by high-speed line scanning. By reasonably adjusting the uniformity of the illumination spot, uneven brightness of the image surface can be effectively suppressed. Uniform image surface light intensity helps to improve the signal-to-noise ratio of abnormal signals and enhances the robustness and accuracy of subsequent algorithm processing.

[0065] If the wafer image does not completely cover the width of the camera target surface (e.g.) Figure 5As shown in (b) above, optimization requires combining a pixel merging strategy with a reasonable scan step size. To prevent aliasing, at least two or three samples are typically acquired perpendicular to the illumination line 404 within each effective optical spot size range; that is, the illumination line width is preferably 2-3 times the scan spacing w. By reasonably configuring the parameters of the line scan camera (such as the merged pixel width h, pixel length d, and scan spacing w), the detection resolution can be further improved while ensuring detection sensitivity. For example, a smaller d value helps improve the resolution in the illumination line direction, a suitable scan spacing w helps optimize the detection accuracy perpendicular to the illumination line direction, and a suitable pixel width h helps improve the signal-to-noise ratio and dynamic range. The above parameters can be flexibly set according to actual detection needs, and the synergistic effect of the three can more accurately identify and detect various anomalies or micro-nano defects on the wafer surface.

[0066] Regardless of whether the wafer image fills the entire width of the camera target area, the Imaging-Motion Modulation Transfer Function (IMMTF) can be used as a unified evaluation standard for the imaging quality of TDI line scan cameras. IMMTF integrates various distortion factors and noise sources related to the optical imaging chain, motion blur during scanning, and pixel width into a single index, facilitating parameter optimization and performance evaluation during the design phase. For the high-speed line scan architecture used in this invention, the IMMTF can be written as:

[0067] ,

[0068] in, Represents the image shift modulation transfer function; The function represents the modulation transfer function of the pure optical imaging link, corresponding to the comprehensive transfer capability of microscope objective 3, tube mirror 4, first objective 8, and second objective 11; Indicates the width of the line spot. This indicates the speed at which the wafer surface moves relative to the line scan direction. This indicates the effective exposure time of a high-speed line scan camera. This indicates the pixel length along the scan direction. This represents the spatial frequency component perpendicular to the line scan direction. This indicates the vertical dimension of a single TDI pixel. This represents the integral series of TDI. This represents the equivalent width of a pixel in the scanning direction. This is used to describe the normalized velocity error of the displacement control component during the integral period, which is also known as the motion synchronization error. The function is used to characterize the frequency response attenuation introduced by finite aperture or finite sampling. It converts the spatial frequency... Take as This means focusing on the system's performance near the Nyquist frequency, which is the most sensitive frequency range for detecting micro- and nano-defects.

[0069] As can be seen from the formula, like the various elements in the MTF... The terms reflect the impact of different physical processes on image contrast. The first term corresponds to motion blur: when the scanning speed... Increase or exposure time When extended, if the width of the line spot... The inability to scale down in tandem results in a loss of contrast in the low to mid-frequency range. By setting a speed-exposure linkage curve in the control system and using a momentum compensation algorithm, this factor can be suppressed to above 0.9, ensuring sufficient low-frequency contrast even during full-frame scanning of a 300mm wafer. The second point reflects the impact of the pixel sampling window: the larger the pixel size, the faster the spatial frequency response decreases. Therefore, this embodiment employs a collaborative strategy of charge domain row merging and digital domain column merging to achieve an effective pixel aspect ratio close to 1:1, thus maintaining the signal-to-noise ratio while minimizing the erosion of resolution.

[0070] The third fractional term takes into account the TDI integral stage, speed error, and pixel drift, and explicitly specifies the integral stage. The product of the speed error and the error itself introduces additional distortion: the higher the TDI level, the more stringent the requirements for motion synchronization. Once... If rapid closed-loop compensation is not possible, significant modulation attenuation will occur at high frequencies. Therefore, we introduced a high-bandwidth interferometric position feedback and feedforward velocity compensation algorithm into the displacement control component, through real-time estimation... And adjust the stage speed so that The first zero of the function is far from the operating frequency band, thus maintaining a high-frequency response of over 60%.

[0071] Furthermore, the image-shifting MTF can be simultaneously optimized with the pixel merging strategy, scan step size, and line spot illumination parameters. By establishing a multivariate objective function based on the above formula, the following parameters are set: With a Nyquist frequency of not less than 0.35 and a half-Nyquist frequency of not less than 0.55, and a signal-to-noise ratio of greater than 45dB, a set of parameters that meet the dual requirements of high speed and high precision can be obtained in the system design stage.

[0072] It is worth emphasizing that the above-mentioned image shift The framework can be used not only for offline design of imaging links but also embedded in online calibration processes. By projecting a standard resolution wafer before wafer mounting or using the built-in fine-wire grid stack, we can measure the image shift of the system in real time. Curve. Once a certain frequency band is detected... If the image quality drops below a threshold, the control system automatically adjusts the exposure, charge accumulation time, line spot energy distribution, or scanning speed, and selects a new pixel merging level using a lookup table algorithm, thus achieving adaptive maintenance without downtime. This dynamic parameter tuning capability ensures that the system maintains stable defect identification capabilities even during long-term online operation and when facing different wafer processes.

[0073] Furthermore, the present invention is not limited to the specific parameter configurations mentioned above. Related imaging methods, pixel merging strategies, and scanning step sizes can all be adjusted and optimized according to different application scenarios, equipment performance, and detection targets to adapt to future processes and diverse detection needs.

[0074] In this embodiment, the pixel merging strategy is as follows: Figure 6 As shown, pixel merging is performed parallel to the scanning direction. Figure 6 (a) in the middle), merged perpendicular to the scanning direction ( Figure 6 (b) in the middle), the two merging methods are combined ( Figure 6 (c) The number of pixel merges parallel to the scanning direction can range from 2 to 256 levels, while the number of pixel merges perpendicular to the scanning direction can range from 2 to 10. This directional pixel merge needs to match the system's optical resolution; the specific matching formula is that the Airy disk radius multiplied by the system magnification is approximately equal to the width of a single pixel. Pixel merge strategies can be charge domain (analog domain) pixel merge, digital domain (signal processing domain) pixel merge, or a combination of both. Charge domain pixel merge helps improve the signal-to-noise ratio but may be limited by dynamic range; digital domain pixel merge offers greater flexibility, but each merge operation may introduce additional readout noise and quantization noise. The specific pixel merge method and number of merge levels should be determined comprehensively based on various factors such as the object being detected, defect type, system noise level, and dynamic range requirements, in order to more accurately identify and detect various anomalies or micro / nano defects on the wafer surface.

[0075] By rationally setting the pixel merging strategy and the size parameters of the merged pixels, an optimal balance between signal-to-noise ratio, resolution, and detection sensitivity can be achieved in different detection scenarios, thereby more accurately identifying and detecting various anomalies or micro / nano defects on the wafer surface. This invention is not limited to the specific merging methods and parameters described above; the relevant pixel merging strategies and their implementation methods can be flexibly adjusted according to actual needs to adapt to the future development of detection technologies and diverse application requirements.

[0076] like Figure 7 As shown, this embodiment of the invention also provides a method for detecting wafer surface defects based on a high-speed line scan camera, including the following steps:

[0077] S1, using an illumination system to generate a line spot of light to illuminate the wafer surface;

[0078] S2, using the wafer scanning component to drive the wafer to move relative to the line spot, to achieve full coverage scanning of the wafer surface;

[0079] S3 uses an imaging system to collect the reflected signals from the wafer surface during the scanning process and image them onto a high-speed line scan camera;

[0080] S4 controls the high-speed line scan camera, which configures the pixel merging mode and level in real time according to the characteristics of the defect to be detected and the detection speed requirements, and acquires and processes the signal to generate a wafer surface image.

[0081] S5 analyzes the generated images to identify defects and optimizes the parameters of the detection system and performs online adaptive calibration through the control system.

[0082] It should be noted that the wafer surface defect detection method based on a high-speed line scan camera provided in the above embodiments belongs to the same inventive concept as the wafer surface defect detection system based on a high-speed line scan camera. For specific implementation details, please refer to the embodiments of the wafer surface defect detection system based on a high-speed line scan camera, which will not be repeated here.

[0083] The specific embodiments described above illustrate the technical solution and beneficial effects of the present invention in detail. It should be understood that the above description is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, additions, and equivalent substitutions made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A wafer surface defect detection system based on a high-speed line scan camera, characterized in that, include: An illumination system for providing bright-field illumination with adjustable illumination angle, and for forming adjustable-size line spots on the wafer surface through a line spot generating device therein; The wafer scanning assembly includes a stage for carrying the wafer and a displacement control assembly for driving the stage to move. The displacement control assembly achieves full-coverage scanning of the wafer surface through a position feedback function. An imaging system, including a bright-field imaging component for collecting reflected signals under bright-field illumination conditions; High-speed line scan camera is used to perform high-speed scanning imaging of signals collected by the imaging system, and balances the signal-to-noise ratio, dynamic range and detection resolution through programmable pixel merging function; The control system coordinates the control of the illumination system, wafer scanning components, imaging system, and high-speed line scan camera, and performs parameter optimization and online adaptive calibration of the detection system based on the image shift modulation transfer function. The parameter optimization of the detection system based on the image shift modulation transfer function includes: constructing a multivariate objective function as the image shift modulation transfer function, which at least includes the optical imaging link, motion blur, pixel sampling effect, TDI integral series, and motion synchronization error, expressed as: ， in, Represents the image-shift modulation transfer function. The function represents the modulation transfer function of a purely optical imaging link. Indicates the width of the line spot. This indicates the speed at which the wafer surface moves relative to the line scan direction. This indicates the effective exposure time of a high-speed line scan camera. This indicates the pixel length along the scan direction. This represents the spatial frequency component perpendicular to the line scan direction. This indicates the vertical dimension of a single TDI pixel. This represents the integral series of TDI. This represents the equivalent width of a pixel in the scanning direction. This is used to describe the normalized velocity error of the displacement control component during the integral period, which is also known as the motion synchronization error. The function is used to characterize the frequency response attenuation introduced by finite aperture or finite sampling. By co-optimizing the above parameters, a parameter combination that meets the preset high speed and high precision requirements can be obtained.

2. The wafer surface defect detection system based on a high-speed line scan camera according to claim 1, characterized in that, The illumination system includes an illumination source for providing an illumination beam, a waveplate for adjusting the polarization state, a first attenuator for adjusting the illumination power, a beam expander for adjusting the beam size and shape, and the line spot generating device arranged sequentially along the optical path. Bright-field illumination of the wafer surface is achieved by controlling the incident angle of the illumination beam.

3. The wafer surface defect detection system based on a high-speed line scan camera according to claim 1 or 2, characterized in that, The line spot generating device under bright field illumination conditions adopts a cylindrical lens group, a diffractive optical element or a combination of the two, or a combination of a separately designed Fourier transform lens and a slit aperture.

4. The wafer surface defect detection system based on a high-speed line scan camera according to claim 1, characterized in that, The displacement control component integrates a high-precision closed-loop control position feedback device to support the execution of rotary spiral or line-by-line serpentine scanning paths.

5. The wafer surface defect detection system based on a high-speed line scan camera according to claim 1, characterized in that, The bright-field imaging assembly includes a microscope objective, a tube mirror, a filter, and a re-imaging assembly. The re-imaging assembly includes a first objective, an analyzer, a second attenuator, and a second objective. In addition, an aperture is provided after the filter in the bright-field imaging assembly to adjust the numerical aperture of the microscope objective.

6. The wafer surface defect detection system based on a high-speed line scan camera according to claim 1, characterized in that, The programmable pixel merging function of the high-speed linear scanner includes merging methods such as merging parallel to the scanning direction, merging perpendicular to the scanning direction, or a combination of both, and merging objects include charge domain pixels, digital domain pixels, or a combination of both.

7. The wafer surface defect detection system based on a high-speed line scan camera according to claim 6, characterized in that, When performing pixel merging perpendicular to the scanning direction, the effective pixel width after merging is matched with the optical resolution of the imaging system. The matching relationship is: the product of the optical Airy disk radius of the imaging system and the magnification of the imaging system is approximately equal to the effective pixel width after merging.

8. The wafer surface defect detection system based on a high-speed line scan camera according to claim 1, characterized in that, Online adaptive calibration of the detection system based on the image shift modulation transfer function includes: During online inspection, the image shift modulation transfer function (EMF) curve of the inspection system is measured in real time or at regular intervals by scanning a known structure on a standard resolution plate or wafer. When the response value of the EMF curve is lower than a preset threshold, the exposure time is automatically adjusted. Scanning speed Linear spot width or pixel equivalent width One or more parameters in the image are adjusted until the response value of the image shift modulation transfer function curve recovers to the target level, thereby achieving dynamic compensation of the detection system performance.

9. A wafer surface defect detection method based on a high-speed line scan camera, implemented using the wafer surface defect detection system based on a high-speed line scan camera as described in any one of claims 1 to 8, characterized in that, Includes the following steps: S1, using an illumination system to generate a line spot of light to illuminate the wafer surface; S2, using the wafer scanning component to drive the wafer to move relative to the line spot, to achieve full coverage scanning of the wafer surface; S3 uses an imaging system to collect the reflected signals from the wafer surface during the scanning process and image them onto a high-speed line scan camera; S4 controls the high-speed line scan camera, which configures the pixel merging mode and level in real time according to the characteristics of the defect to be detected and the detection speed requirements, and acquires and processes the signal to generate a wafer surface image. S5 analyzes the generated images to identify defects and optimizes the parameters of the detection system and performs online adaptive calibration through the control system.

Images