Wafer surface defect detection system and detection method

Abstract

The invention relates to the technical field of wafer detection, in particular to a wafer surface defect detection system and method.The system comprises an imaging module, the imaging module comprises at least two sets of imaging channels, each imaging channel comprises an imaging camera for collecting light signals and converting the light signals into image signals, and a light modulation assembly is arranged in each imaging channel; the light modulation assembly performs polarization modulation and light intensity modulation on emergent light reflected or scattered by the surface of the wafer; the polarization control module adjusts the rotation angle of the light modulation assembly in each imaging channel; the signal processing module carries out differential operation on different image signals, light rays emitted by the multiple groups of imaging channels are set with different polarization parameters and light intensity parameters through the light modulation assembly, multiple groups of image signals are obtained, and an imaging image about scattered light of the same detection area is formed. The imaging image comprises an image with the defect signal as weak as possible and an image with the defect signal as strong as possible, and the imaging image with the defect signal enhanced can be obtained through differential operation.

Description

Technical Field

[0001] This invention relates to the field of wafer inspection technology, and in particular to a wafer surface defect detection system and method. Background Technology

[0002] In the semiconductor manufacturing industry, wafer quality control is crucial, and defect detection is a key aspect of quality control. Wafer defects are diverse, and their presence significantly impacts wafer yield and economic efficiency. Therefore, various wafer inspection methods have emerged in the market. Among them, dark field wafer inspection technology, with its advantages of high detection speed and high sensitivity, is widely used in wafer defect detection. It uses an illumination system to illuminate the wafer, and an imaging system collects the scattered light from the defects to identify them.

[0003] However, as chip integration density continues to increase, manufacturing processes place stringent demands on wafer defect detection equipment to detect smaller defects and achieve higher detection rates. In this context, light scattering on nanometer- or sub-nanometer-scale undulations on the sample surface generates background noise. When the defect size is too small, the detection signal representing the wafer defect is easily drowned out by the background noise. Although switching the polarization state configuration of the illumination and imaging system can enhance the detection signal of a certain type of defect, it is impossible to simultaneously enhance the detection signals of multiple different defects. Summary of the Invention

[0004] To address the problems of low signal-to-noise ratio in small defect imaging and inability to simultaneously enhance the detection signals of multiple different small defects in existing detection systems, this invention provides a wafer surface defect detection system and method.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a wafer surface defect detection system, comprising an imaging module, the imaging module comprising at least two imaging channels, the imaging channel comprising an imaging camera for acquiring optical signals and converting the optical signals into image signals, and an optical modulation component disposed within the imaging channel, the optical modulation component being used to perform polarization modulation and intensity modulation on the outgoing light reflected or scattered by the wafer surface; A polarization control module is used to independently adjust the rotation angle of the light modulation component in each imaging channel according to a preset angle group. The signal processing module is connected to the imaging module. The signal processing module is used to acquire the image signal corresponding to each imaging channel and perform differential operation on the different image signals to acquire the imaging image.

[0006] Preferably, the imaging channel includes a first imaging channel and a second imaging channel, the preset angle group includes a first angle and a second angle, the image signal includes a defect signal, the optical modulation component includes a polarization modulation component, and the polarization control module is used to adjust the polarization modulation component in the first imaging channel to the first angle where the defect signal reaches its maximum value, and to adjust the polarization modulation component in the second imaging channel to the second angle where the defect signal reaches its minimum value.

[0007] Preferably, the light modulation component includes a light intensity modulation component, the light intensity modulation component includes at least two sets of light attenuation elements with different attenuation rates and at least one light-transmitting channel without the light attenuation elements, and the imaging module includes a switching device, the light intensity modulation component being mounted on the switching device.

[0008] Preferably, the detection system includes at least one side-channel imaging module and a center-channel imaging module, defining the direction perpendicular to the wafer surface as the wafer normal. The center-channel imaging module collects the outgoing light reflected or scattered from the wafer normal, and the side-channel imaging module collects the outgoing light at an angle to the wafer normal.

[0009] Preferably, the system includes an illumination module for providing illumination light to irradiate the wafer to be inspected, the illumination light including at least one of parallel polarized light, vertically polarized light, or elliptically polarized light.

[0010] To solve the above-mentioned technical problems, the present invention provides another technical solution as follows: a method for detecting wafer surface defects, applied to the wafer surface defect detection system as described above, the method comprising the following steps: providing at least two imaging channels, acquiring optical signals within the imaging channels, and converting the optical signals into image signals, wherein an optical modulation component is provided within the imaging channels; The rotation angle of the light modulation component in each imaging channel can be independently adjusted using a preset angle group; At least two sets of image signals corresponding to each of the imaging channels are acquired, and differential operations are performed on the at least two sets of image signals to acquire an imaging image.

[0011] Preferably, the imaging channel includes a first imaging channel and a second imaging channel, and the preset angle group includes a first angle and a second angle. Adjusting the rotation angle of the optical modulation component in each imaging channel independently according to the preset angle group includes the following steps: adjusting the optical modulation component in the first imaging channel to the first angle where the defect signal reaches its maximum value, and adjusting the optical modulation component in the second imaging channel to the second angle where the defect signal reaches its minimum value.

[0012] Preferably, the image signal includes a first image signal acquired based on the first imaging channel and a second image signal acquired based on the second imaging channel. Adjusting the rotation angle of the light modulation component in each imaging channel independently with a preset angle group further includes the following steps: adjusting the light intensity of the first imaging channel and the second imaging channel so that the background noise intensity of the first image signal and the second image signal is consistent.

[0013] Preferably, acquiring at least two sets of image signals corresponding to each of the imaging channels, performing differential operations on the at least two sets of image signals to acquire an imaging image further includes the following steps: analyzing and calculating the signal-to-noise ratio of the at least two sets of image signals to obtain the differential signal-to-noise ratio of different image signals.

[0014] Preferably, the method further includes the following steps: traversing the rotation angle combinations of the optical modulation components in at least two groups of imaging channels, calculating the differential signal-to-noise ratio under each angle combination, and obtaining angle combinations greater than a preset threshold.

[0015] Compared with the prior art, the wafer surface defect detection system and method provided by the present invention have the following beneficial effects: 1. The detection system provided in this embodiment of the invention uses an optical modulation component to modulate the polarization state and intensity of light reflected or scattered from the wafer surface, capturing the modulated polarized light to identify defects in the wafer. For smaller defects, the weak signal may mix with background noise. The imaging module sets at least two imaging channels, and a polarization control module sets different polarization control parameters and intensity parameters for the optical modulation component in each imaging channel. This allows multiple sets of image signals about the same area to be acquired simultaneously through different imaging channels, forming multiple polarized images and providing multi-dimensional data information for subsequent small defect extraction. The signal processing module performs differential operations on the image signals about the same area from different imaging channels, eliminating background noise such as speckle and haze, retaining only the difference signal caused by small defects. By combining the differential results of multiple image signals about the same area within multiple imaging channels, the features of small defects can be reconstructed, and other information about the small defects can be obtained. Therefore, the detection system provided by this invention, through the collaborative design of multiple imaging channels and multi-dimensional polarization modulation, is significantly superior to single-channel intensity imaging technology. Compared with the limitation of a single channel being able to detect only one type of small defect, the multiple imaging channels of this system can simultaneously perform parallel detection of multiple types of small defects. With the dynamic control of light polarization state and light intensity parameters by the polarization modulation component, the simultaneous acquisition and fusion calculation of multiple imaging channels improves the sensitivity and reliability of different types of defect detection. In terms of detection efficiency, the parallel operation of multiple imaging channels within the same imaging module changes the inefficient mechanism of single-channel multi-defect scanning, greatly shortening the detection time of multiple types of defects and adapting to the detection needs of different types of defects in wafer mass production scenarios.

[0016] 2. In this embodiment of the invention, the polarization control module dynamically and independently adjusts the light modulation components in the first and second imaging channels to a first angle and a second angle, thereby optimizing the polarization state of the light according to the characteristics of the defect. This adjusts the angle to the maximum or minimum value of the defect signal obtained by the signal processing module, significantly improving the contrast between the defect and the background. The polarization control module fixes the optimal polarization angle, reducing interference from ambient polarized light on the random polarization components of the detection system, making the detection results more stable and reducing the false detection rate. When the light modulation component is adjusted to the first angle where the defect signal is at its maximum value, the first imaging channel has the strongest response to the polarized light of the defect, resulting in a more accurate image of the defect. The brightness or grayscale value of the defect differs from the rest of the area. The defect appears brightest in the imaging image. When the light modulation component is adjusted to the second angle where the defect signal is at its minimum, the defect signal is suppressed to its minimum value. The imaging of the defect signal at this angle is weakened, and the defect is not displayed or has the weakest brightness in the imaging image. At this time, the imaging image of the second imaging channel can be used as a reference image to remove the interference of background noise. Therefore, by performing a differential operation with the imaging image of the first imaging channel, the common signals of background noise can be further eliminated, and only the characteristic differences of the defect are retained, thereby improving the sensitivity of defect detection. It is suitable for the detection of small defects in complex backgrounds.

[0017] 3. In this embodiment of the invention, speckle and haze noise are noises with no or weak polarization characteristics. Changes in the parameters of the optical modulation component have little impact on the imaging intensity of speckle and haze noise or are asynchronous with the small defect polarization imaging signal. For noises with polarization characteristics such as speckle and haze noise, the light intensity in different imaging channels is adjusted by different light transmittance of the optical attenuation element, so that the brightness of speckle and haze noise is consistent in different imaging channels, avoiding the impact of such polarization noise on detection accuracy. At least two sets of optical attenuation elements with different light transmittance continuously or gradiently adjust the light intensity in the imaging channel to adapt to the defect scattering characteristics and detection accuracy of different wafer surfaces. The through-hole without optical attenuation element has 100% light transmittance and is used when the strongest light signal is required, ensuring that the imaging module receives a sufficiently strong image signal and avoiding image blurring or information loss due to insufficient light intensity. The switching device is used to quickly switch the optical intensity modulation component to a no-attenuation state or different attenuation states, reducing the complexity of the detection system, reducing optical path offset error, and ensuring the stability and repeatability of optical intensity modulation.

[0018] 4. In this embodiment of the invention, the central imaging module collects reflected or scattered light in the direction perpendicular to the wafer surface, i.e., the normal direction, mainly detecting defects perpendicular or nearly perpendicular to the wafer surface, such as surface protrusions, depressions, and vertical cracks. These defects show significant changes in the intensity of scattered light in the normal direction. At the same time, the central imaging module provides a coordinate reference for the side imaging modules, facilitating the calibration of spatial positions for multiple imaging channels. The side imaging modules collect large-angle emitted light at a certain angle to the normal of the wafer surface, detecting defects with tilt angles, such as sidewall damage, oblique cracks, and edge foreign objects. The central and side imaging modules comprehensively detect defects from multiple angles. When only one imaging module works, such as the central module, the system can maintain basic defect detection. When two or three imaging modules work simultaneously, they work together to improve detection accuracy through multi-source data fusion.

[0019] 5. In this embodiment of the invention, during wafer inspection, the illumination module controls the polarization state of the illumination light to be at least one of parallel polarized light, vertical polarized light, or elliptically polarized light, thereby adjusting the interaction between the illumination light and the wafer surface structure, thus optimizing the contrast, sensitivity, and defect identification capability of the detection signal. For example, when parallel polarized light and vertical polarized light are used for simultaneous illumination, data signals of surface flatness and defect contrast can be acquired simultaneously, improving the accuracy of defect classification.

[0020] 6. This invention also provides a detection method applied to the above-mentioned detection system. In this method, the angle of the light modulation component is changed in at least two imaging channels. The polarization characteristics of background noise, such as speckle and haze noise, are usually consistent. However, different defects, such as scratches, particles, dirt, bridging, short circuits, and subsurface stacking faults and dislocations, have a greater polarization response to light than a light intensity response. The light emitted after reflection or scattering from the wafer surface contains different light signals under different polarization parameters. The emitted light in each imaging channel forms different imaging images based on different polarization angles. In at least one imaging channel, an image with a significant contrast between the defect and background noise and a strong defect signal is obtained. In another imaging channel, an image with a weak defect signal is obtained. After differential operation, the background noise is canceled out, and the image signal of the defect is highlighted. In this method, polarization differential imaging of at least two imaging channels breaks through the limitations of detection with a single imaging channel. It uses the difference in polarization response rather than the difference in light intensity to identify defects and suppresses common background noise in the same area through differential operation, adapting to complex detection environments.

[0021] 7. In this embodiment of the invention, a single imaging channel can only acquire signals at a fixed polarization angle. Defect identification relies on the intensity or phase change of light under a single polarization state, which is easily affected by background noise and cannot distinguish between defects and related noise. However, the multiple imaging channels in this invention focus on the polarization change caused by the defect and suppress the background polarized light, highlighting the polarization difference between the defect and the background noise. The first imaging channel and the second imaging channel obtain the signal difference between the maximum and minimum values ​​of the defect signal through the angle of the modulation light modulation component. After differential operation by the signal processor, the background noise is not obvious or is not displayed at all, and the defect signal can be separated from the background noise.

[0022] 8. In this embodiment of the invention, the difference in adjustment of defects and background noise by the optical modulation components in different imaging channels is utilized. The intensity of background noise is made consistent in the two imaging channels by adjusting the optical intensity. Then, the background noise is canceled out by signal subtraction, and the defect signal is retained. The background noise includes speckle and haze noise. Speckle is random bright and dark spots generated by laser interference. Haze noise is diffuse scattering background caused by small undulations or impurities on the wafer surface. When the intensity of the two imaging channels is adjusted to make the background noise consistent, the speckle and haze noise will cancel each other out in the differential operation due to the consistent distribution in different channels of the same imaging module.

[0023] 9. In this embodiment of the invention, the signal-to-noise ratio (SNR) is the ratio of the target signal power to the background noise power in the image signal. The differential SNR refers to the difference in SNR between the differential signals of the images in at least two imaging channels. The differential SNR can be set to a threshold. When the differential SNR obtained after the optical modulation component rotates by an angle exceeds the threshold, it can be determined as a better setting angle, i.e., a preset angle group. In this invention, the differential SNR is upgraded from the single SNR of a single imaging channel to the differential SNR difference assessment between multiple imaging channels through comparative analysis. The ability to identify defects is improved by the differential SNR difference characteristics, and a quantitative comparison basis is provided for multiple sets of data obtained from multiple imaging channels.

[0024] 10. In this embodiment of the invention, by traversing all possible combinations of rotation angles of the optical modulation components, the optimal setting angle is found when the differential signal-to-noise ratio exceeds a threshold. For example, when the angle of the optical modulation component in the first imaging channel is A, the angle of the optical modulation component in the second imaging channel is rotated 360° and the differential signal-to-noise ratio with that of the image signal in the first imaging channel is calculated. When the angle of the optical modulation component in the second imaging channel is B, the differential signal-to-noise ratio reaches or exceeds the preset differential signal-to-noise ratio threshold. The defect signals of the image signals in the first and second imaging channels are significantly amplified through differential operation, while background noise such as speckle and fog noise is suppressed. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a block diagram of the detection system provided in the first embodiment of the present invention.

[0027] Figure 2 This is a block diagram of the signal processing module of the detection system provided in the first embodiment of the present invention.

[0028] Figure 3 This is a block diagram of the imaging module of the detection system provided in the first embodiment of the present invention.

[0029] Figure 4 This is a block diagram of the imaging channel of the detection system provided in the first embodiment of the present invention.

[0030] Figure 5 This is a schematic diagram of the optical path of the detection system provided in the first embodiment of the present invention.

[0031] Figure 6 This is a schematic diagram of the optical path of the second illumination module of the detection system provided in the first embodiment of the present invention.

[0032] Figure 7 This is a schematic diagram of the optical path of the imaging channel and signal processing module of the detection system provided in the first embodiment of the present invention.

[0033] Figure 8 This is a schematic diagram of the structure of the light intensity modulation component of the detection system provided in the first embodiment of the present invention.

[0034] Figure 9 This is an example diagram of the detection results of the imaging module of the detection system provided in the first embodiment of the present invention.

[0035] Figure 10 This is a flowchart of the detection method provided in the second embodiment of the present invention.

[0036] Explanation of reference numerals in the attached diagram: 100. Detection system; 1. Imaging module; 2. Polarization control module; 3. Signal processing module; 4. Illumination module; 5. Second illumination module; 6. Central imaging module; 7. Side imaging module; 8. Wafer; 10. Imaging channel; 11. Imaging camera; 12. Optical modulation assembly; 13. Imaging objective lens; 14. Mirror; 15. Beam splitter; 16. Imaging tube lens; 31. Image comparator; 41. Laser; 42. Polarizer; 43. Quarter wave plate; 44. Half wave plate; 51. Main laser; 52. Secondary laser; 53. Polarizer; 54. Second half wave plate; 55. Beam combiner prism; 71. Left imaging module; 72. Right imaging module; 101. First imaging channel; 102. Second imaging channel; 121. Polarization modulation component; 122. Light intensity modulation component; 1221. Optical attenuation element; 1222. Light transmission channel; 1223. Switching device. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0038] In the embodiments provided by this invention, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0039] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of the invention. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Those skilled in the art should also recognize that the embodiments described in the specification are optional embodiments, and the actions and modules involved are not necessarily essential to the invention.

[0040] In various embodiments of the present invention, it should be understood that the sequence number of each process does not necessarily imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0041] The flowcharts and block diagrams in the accompanying drawings illustrate methods and possible architectures, functions, and operations according to various embodiments of this application. In this regard, each block in the flowchart or block diagram may represent part of a step. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending on the functions involved.

[0042] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0043] In the embodiments provided by this invention, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0044] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of the invention. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Those skilled in the art should also recognize that the embodiments described in the specification are optional embodiments, and the actions and modules involved are not necessarily essential to the invention.

[0045] In various embodiments of the present invention, it should be understood that the sequence number of each process does not necessarily imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0046] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending on the functions involved. It is particularly important to note that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0047] Please see Figures 1-4 The first embodiment of the present invention provides a detection system 100 for surface defects of wafer 8. The detection system 100 includes: an imaging module 1, the imaging module 1 includes at least two imaging channels 10, the imaging channel 10 includes an imaging camera 11 that collects light signals and converts the light signals into image signals, and an optical modulation component 12 is provided in the imaging channel 10. The optical modulation component 12 is used to perform polarization modulation and light intensity modulation on the outgoing light reflected or scattered by the surface of wafer 8. Polarization control module 2 is used to independently adjust the rotation angle of the light modulation component 12 in each imaging channel 10 according to a preset angle group. Signal processing module 3 is connected to imaging module 1. Signal processing module 3 is used to acquire the image signals corresponding to each imaging channel 10, perform differential operations on different image signals, and acquire imaging images.

[0048] Understandably, the wafer 8 surface defect detection system 100 provided by the present invention is an optical detection system 100, which uses laser as incident light to irradiate the wafer 8 surface at a specific angle. When there is a defect on the wafer 8 surface, the light will be scattered at the defect, while in the defect-free area, the light will not be scattered or will be scattered weakly. The imaging module 1 is located at a position with a different incident angle from the illumination light, collects the scattered light generated by the defect, and converts it into an electrical signal. By analyzing and processing the electrical signal to form an image, the location, size and other information of the wafer 8 surface defect can be determined.

[0049] Specifically, the scattered light from defects on the surface of wafer 8 differs from that from the scattered light from a normal surface in polarization. The polarization control module 2 controls the light modulation components 12 in different imaging channels 10 to modulate the polarization state of the light scattered from the surface of wafer 8, selectively allowing the scattered light from the target defect to pass through or be suppressed, thereby enhancing the contrast between the features of the target defect and the background noise.

[0050] As chip integration increases, the manufacturing process demands that the wafer defect detection system 100 detect smaller defects and achieve higher detection rates. Nanoscale or sub-nanometer undulations and subsurface defects on the wafer 8 surface, like the target defects to be detected, will also scatter under incident light, thus forming stray light that interferes with the detection of the target defects. When the target defect is smaller than a certain size, the intensity of the stray light will be close to or even greater than the signal light intensity of the corresponding target defect, causing the detection signal of the target defect to be submerged by background noise.

[0051] In the detection system 100 provided by the present invention, the imaging module 1 includes at least two imaging channels 10. The multiple imaging channels 10 can set different polarization parameters and light intensity parameters for the light incident into the imaging channel 10 in each imaging channel 10, and simultaneously acquire multiple sets of image signals of the same area from different imaging channels 10, forming imaging images of the scattered light of the same detection area with different polarization parameters. The imaging images with different polarization parameters include images with the weakest possible defect signal and images with the strongest possible defect signal after polarization state modulation and light intensity modulation.

[0052] Understandably, the detection system 100 provided by the present invention, through the collaborative design of multiple imaging channels 10 and multidimensional polarization modulation, is significantly superior to single-channel intensity imaging technology. Compared with the limitation of a single channel being able to detect only one type of small defect, the multiple imaging channels 10 of this system can simultaneously perform parallel detection of multiple types of small defects. With the dynamic control of light polarization state and light intensity parameters by the polarization modulation component 121, the multiple imaging channels 10 can simultaneously acquire and fuse calculations, improving the sensitivity and reliability of different types of defect detection. In terms of detection efficiency, the multiple imaging channels 10 within the same imaging module 1 work in parallel, changing the inefficient mechanism of single-channel multi-defect scanning, and significantly shortening the detection time of multiple types of defects, adapting to the detection needs of different types of defects in wafer mass production scenarios.

[0053] As a non-limiting embodiment, in this embodiment, the light modulation component 12 includes a polarization modulation component 121. The polarization modulation component 121 can control the polarization direction of the light entering the imaging module 1, reduce background noise and stray light interference, and improve the contrast and clarity of the image. The polarization modulation component 121 can be an analyzer or a polarization grating, etc., and is not specifically limited in this invention, as long as it can achieve the function of changing and controlling the polarization parameters of the light.

[0054] When the polarization direction of the scattered light from the target defect is aligned with the optical axis of the polarization modulation component 121, the defect signal is enhanced; when the polarization direction of the scattered light from the background noise is perpendicular to the transmission axis of the polarization modulation component 121, the image signal corresponding to the background noise is weakened.

[0055] The signal processing module 3 performs differential operations on the image signals of the same area from different imaging channels 10, which can eliminate background noise with consistent background, such as speckle and haze noise, and retain only the difference signals caused by small defects. By combining the differential results of image signals from multiple imaging channels 10, imaging images of various small defects can be obtained.

[0056] Compared to single-channel two-dimensional intensity imaging, the detection system 100 in this invention performs differential fusion calculation on image information data about the same area in multiple imaging channels 10, further improving the detection accuracy of the detection system 100. Moreover, compared to single-channel multi-defect scanning, multi-channel simultaneous imaging significantly shortens the detection time for different types of defects on the same wafer 8, and can adapt to the detection needs of different types of defects in the mass production of wafer 8. Furthermore, existing polarization excitation and detection optimization methods have limitations in improving the signal-to-noise ratio.

[0057] Furthermore, the signal processing module 3 includes an image comparator 31 connected to at least two imaging channels 10. The image comparator 31 is used to compare image data from different imaging channels 10, identify and analyze differences. In this embodiment, the image comparator 31 performs differential operations on the image signals of the two imaging channels 10 to suppress background noise, highlight the target defect signal intensity, and improve the defect detection signal-to-noise ratio.

[0058] Furthermore, an imaging camera 11 is provided in the imaging channel 10. The imaging camera 11 is used to receive and record the light signal modulated by the light modulation component 12. The imaging camera 11 is located at the end of each imaging channel 10 and converts the focused light signal into an electrical signal for image recording and processing.

[0059] The imaging camera 11 in each imaging channel 10 is located at the end of the imaging channel 10. The imaging camera 11 is connected to the image comparator 31 by cable / fiber. The image comparator 31 receives image signals from different imaging cameras 11 and compares and analyzes the different image signals through algorithms.

[0060] Specifically, the imaging camera 11 can be a time delay integration (TDI) camera, which increases the exposure time through delay integration, thereby collecting more light signals and improving the sensitivity of imaging.

[0061] Optionally, the imaging camera 11 can also be a charge-coupled device (CCD) camera or a complementary metal-oxide-semiconductor (CMOS) camera. In this embodiment, no specific limitation is made, as long as it can realize the function of converting the light signal gathered by the imaging objective lens 13 into an electrical signal and transmitting it to the image comparator 31.

[0062] Furthermore, the imaging module 1 is also equipped with a reflector 14, a beam splitter 15, an imaging tube lens 16, and an imaging objective lens 13, which are shared by multiple imaging channels 10.

[0063] Understandably, the reflector 14 is used to change the direction of light so that the light can correctly enter the imaging channel 10. It is positioned between the imaging tube lens 16 and the light modulation component 12. The beam splitter 15 is used to split the outgoing light reflected or scattered by the surface of the wafer 8 into two or more beams, which enter different imaging channels 10 respectively. It is positioned at the beginning of the imaging channel 10 and is coaxially arranged with the imaging tube lens 16 and the imaging objective lens 13. The imaging tube lens 16 is used to adjust the focal length and optical path of the imaging module 1 to ensure clear imaging. The imaging objective lens 13 focuses the image of the object onto the photosensitive surface of the imaging camera 11, providing a basis for image formation. It is positioned at one end of the imaging module 1 near the surface of the wafer 8. These components, together with other components in the imaging channel 10, constitute a complete imaging module 1, which can detect and analyze the polarization characteristics of the object.

[0064] As shown as a non-limiting embodiment, such as Figure 7 The imaging module 1 shown includes two imaging channels 10, such as... Figure 5 The imaging module 1 shown includes three imaging channels 10.

[0065] Furthermore, the signal processing module 3 is also used to analyze and calculate the signal-to-noise ratio of at least two sets of image signals to obtain the differential signal-to-noise ratio of different image signals.

[0066] Specifically, the signal processing module 3 first acquires at least two sets of image signals from different imaging channels 10 of the imaging module 1. These image signals include signal data of the target defect and signal data of background noise. The signal processing module 3 aligns and subtracts the two sets of image signals to calculate the signal data of the target defect and the signal data of background noise in the differential image signal, i.e., the differential signal-to-noise ratio.

[0067] Understandably, differential signal-to-noise ratio is used to compare image signals acquired by different imaging channels 10 to help select the best imaging parameters. By comparing the signal-to-noise ratio under different polarization parameters of scattered light, the rotation angle of the polarization modulation component 121 that is most sensitive to the target defect can be determined, and the most suitable rotation angle can be selected for imaging, thereby improving the accuracy and reliability of detection.

[0068] Specifically, please refer to Figure 4 The imaging channel 10 includes a first imaging channel 101 and a second imaging channel 102. The preset angle group includes a first angle and a second angle. The image signal includes a defect signal. The optical modulation component 12 includes a polarization modulation component 121. The image signal includes a first image signal acquired based on the first imaging channel 101 and a second image signal acquired based on the second imaging channel 102.

[0069] In one non-limiting embodiment, the preset angle group includes a first angle that adjusts the polarization modulation component 121 in the first imaging channel 101 to the maximum value of the defect signal, and / or a second angle that adjusts the polarization modulation component 121 in the second imaging channel 102 to the minimum value of the defect signal, and / or a combination of rotation angles that maximizes the differential signal-to-noise ratio of the first image signal acquired based on the first imaging channel 101 and the second image signal acquired based on the second imaging channel 102, and that the maximum value exceeds a preset threshold.

[0070] Furthermore, the polarization control module 2 is used to adjust the polarization modulation component 121 in the first imaging channel 101 to a first angle where the defect signal reaches its maximum value, and to adjust the polarization modulation component 121 in the second imaging channel 102 to a second angle where the defect signal reaches its minimum value.

[0071] Understandably, imaging channel 10 is used to acquire image signals of scattered light from the surface of wafer 8. The defect signal is adjusted to the maximum and minimum values ​​in the two imaging channels 10 respectively. Combined with image differential operation, the differential signal-to-noise ratio between the defect signal and the background noise can be enhanced, thereby achieving high-sensitivity detection of minute defects.

[0072] Furthermore, the image signals acquired by the imaging channel 10 include a first image signal acquired through the first imaging channel 101 and a second image signal acquired through the second imaging channel 102, wherein both the first image signal and the second image signal contain defect signals targeting the same defect.

[0073] The polarization control module 2 adjusts the angle of the polarization modulation component 121 in the first imaging channel 101 to the first angle at which the defect signal in the first image signal acquired by the first imaging channel 101 reaches its maximum value, and adjusts the angle of the polarization modulation component 121 in the second imaging channel 102 to the second angle at which the defect signal reaches its minimum value.

[0074] In another embodiment, the preset angle group is a combination of rotation angles that maximizes the differential signal-to-noise ratio between the first image signal and the second image signal, and that the maximum value exceeds a preset threshold.

[0075] The differential signal-to-noise ratio is the signal-to-noise ratio of the imaging difference between the two imaging channels 10. The preset angle group is obtained by traversing and searching the rotation angle combination of the optical modulation components 12 in at least two imaging channels 10. That is, the preset angle group is to make the rotation angle of the polarization modulation component 121 in the first imaging channel 101 and the rotation angle of the polarization modulation component 121 in the second imaging channel 102 obtained by traversing and searching.

[0076] For example, the traversal search execution process for the rotation angle of the polarization modulation component 121 in the first imaging channel 101 and the rotation angle of the polarization modulation component 121 in the second imaging channel 102 is as follows: First: Set the angle of the polarization modulation component 121 in the first imaging channel 101 to an initial value, which can be 0° or any preset initial value. Adjust the angle of the polarization modulation component 121 in the second imaging channel 102 to perform a scan from 0° to 360°. Record the differential signal-to-noise ratio (SNR) with the largest difference between the first and second image signals under this initial value. Second: Increment the angle of the polarization modulation component 121 in the first imaging channel 101. The increment range can be between 0.1° and 1°, which is set by those skilled in the art based on experience. Repeat the above steps for the polarization modulation component 121 in the second imaging channel 102 to obtain an updated maximum differential SNR. Finally: Iterate continuously until the angle of the polarization modulation component 121 in the first imaging channel 101 has traversed the range from 0° to 360°, generating a differential SNR matrix of 3600×3600 angle combinations (0.1° increments).

[0077] Optionally, during the angle adjustment iteration process, the angle combination that reaches the maximum value of differential signal-to-noise ratio and that the maximum value exceeds a preset threshold can be found. Specifically, in this embodiment, the preset threshold is set to 4, that is, the process can be stopped when the corresponding angle where the differential signal-to-noise ratio between different imaging channels 10 reaches 4 is found.

[0078] Specifically, the polarization control module 2 can achieve angle control of the optical modulation component 12 by a high-precision rotator. The rotator includes a stepper motor, an angle sensor and a control circuit board to ensure the acquisition of the optimal defect signal modulation effect and realize the automated operation of angle adjustment, improve detection efficiency, reduce manual intervention, and is suitable for large-scale mass production inspection of semiconductor wafers 8.

[0079] As a non-limiting specific embodiment, in the detection system 100 provided by the present invention, the rotation range of the rotator driving the optical modulation component 12 is 0-360°, with a minimum step size of 0.1°.

[0080] Specifically, the beam splitter 15 splits the outgoing light reflected or scattered by the surface of the wafer 8 into two or more beams, which enter different imaging channels 10 respectively. Each imaging channel 10 includes a corresponding optical axis, and the rotator drives the polarization modulation component 121 to rotate around the optical axis in its respective imaging channel 10.

[0081] Further, see Figure 8 The light modulation component 12 includes a light intensity modulation component 122, which includes at least two sets of light attenuation elements 1221 with different light transmittance and at least one light transmission channel 1222 without light attenuation elements 1221. The imaging module 1 includes a switching device 1223, and the light intensity modulation component 122 is mounted on the switching device 1223.

[0082] Understandably, in the actual manufacturing process, background noise includes speckle and haze noise. Speckle is random bright and dark spots generated by laser interference, while haze noise is diffuse scattering background caused by small undulations or impurities on the surface of wafer 8. When the light intensity of the two imaging channels 10 is adjusted to match the background noise, the speckle and haze noise will cancel each other out in the subtraction operation due to their consistent distribution in different imaging channels 10 of the same imaging module 1.

[0083] Speckle and fog noise are non-polarized or weakly polarized imaging. The polarization control module 2 rotates the light modulation component 12, which has little impact on the imaging intensity of speckle and fog noise. For noise with polarization characteristics such as speckle and fog noise, the imaging intensity is adjusted by an attenuator to make its brightness consistent in different imaging channels 10, so as to avoid excessive differences in this type of noise from affecting the detection accuracy.

[0084] Specifically, the light intensity modulation component 122 is composed of a series of light attenuation elements 1221 with proportionally varying light transmittance installed in the switching device 1223. The switching device 1223 includes at least one light transmission channel 1222 without light attenuation elements 1221. The light intensity modulation component 122 may include one switching device 1223 or multiple switching devices 1223. The switching device 1223 is coaxially arranged with the optical axis of the imaging channel 10. By switching the light attenuation elements 1221 in the imaging channel 10, imaging with different brightness can be achieved.

[0085] As shown in the figure, the light intensity modulation component 122 is coaxially arranged with the polarization modulation component 121. The light intensity modulation component 122 is a circular turntable with 12 attenuator mounting positions evenly distributed on the turntable. Each mounting position corresponds to a light attenuation element 1221 with a different attenuation coefficient. By switching the light attenuation element 1221 with different attenuation coefficients through the switching device 1223, flexible control of light intensity can be achieved.

[0086] Furthermore, the system includes an illumination module 4, which provides illumination light to illuminate the wafer 8 to be inspected. The illumination light includes at least one of parallel polarized light, vertical polarized light, or elliptically polarized light.

[0087] Understandably, by providing illumination light with different polarization states—parallel polarized light, vertical polarized light, or elliptically polarized light—the scattering signal of specific types of defects can be enhanced, while background noise can be suppressed, thereby improving detection sensitivity and resolution.

[0088] Specifically, the illumination module 4 includes a laser 41, a polarizer 42, a quarter-wave plate 43, a half-wave plate 44, and a reflector 14. The quarter-wave plate 43 can rotate around the optical axis inside the illumination module 4, and can also switch to enter or exit the optical path of the illumination module 4. The half-wave plate 44 can rotate around the optical axis inside the illumination module 4, and can also switch to enter or exit the optical path.

[0089] For example, the processes for generating illumination light with different polarization states are shown below: Laser 41 outputs laser light, which is then polarized by polarizer 42 to generate linearly polarized light. The quarter-wave plate 43 and half-wave plate 44 are controlled to exit the optical path, leaving only the linearly polarized light output by polarizer 42. At this time, the irradiation light is parallel polarized light, which is used to irradiate wafer 8.

[0090] Laser 41 outputs laser light, which is polarized by polarizer 42 to generate linearly polarized light. Half-wave plate 44 is inserted into the optical path and rotated so that the angle between half-wave plate 44 and the direction of linearly polarized light output by polarizer 42 is 45°. At this time, the polarization direction of the outgoing light is rotated by 90° to form vertically polarized light, which is then reflected by mirror 14 to illuminate wafer 8.

[0091] Laser 41 outputs laser light, which is polarized by polarizer 42 to generate linearly polarized light. A quarter-wave plate 43 is inserted into the optical path, and the angle of the quarter-wave plate 43 is adjusted by rotation. After passing through the quarter-wave plate 43, the linearly polarized light is elliptically polarized light, which is then reflected by mirror 14 onto the wafer 8.

[0092] As a non-limiting embodiment, please refer to Figure 6 An alternative to illumination module 4 is a second illumination module 5, which includes a main laser 51, a secondary laser 52, two polarizers 53, a second half-wave plate 54, and a beam combiner prism 55.

[0093] Specifically, the second illumination module 5 adopts a dual-source beam combining optical path design. The main laser 51 and the polarizer 53 generate pure linearly polarized light (P) in a single direction. The auxiliary laser 52, the polarizer 53 and the second half-wave plate 54 generate linearly polarized light (S). The beam combining prism 55 superimposes the linearly polarized light (P) and the linearly polarized light (S) to output a combination of mixed polarized light (P+S).

[0094] Furthermore, the detection system 100 includes at least one side-path imaging module 7 and a center-path imaging module 6. The direction perpendicular to the surface of the wafer 8 is defined as the wafer 8 normal. The center-path imaging module 6 collects the outgoing light reflected or scattered from the wafer 8 normal, and the side-path imaging module 7 collects the outgoing light that forms an angle with the wafer 8 normal.

[0095] Understandably, the detection system 100 includes at least one side-path imaging module 7, which can be expanded to left and right side-path imaging modules 7. The side-path imaging module 7 includes a left-path imaging module 71 and a right-path imaging module 72. The left-path imaging module 71 and the right-path imaging module 72 are symmetrically distributed on both sides of the central imaging module 6, forming a preset angle with the normal of the wafer 8. In the spatial dimension, the direction perpendicular to the surface of the wafer 8 is defined as the normal of the wafer 8, that is, the direction forming a 90° angle with the surface of the wafer 8, which serves as the spatial reference benchmark of the optical detection system 100. The left-path imaging module 71 and the right-path imaging module 72 work together to symmetrically collect large-angle scattered light from both sides of the wafer 8, detect large-scattering-angle defects, and at the same time combine the central detection of a certain type of defect in a specific direction, so as to accurately measure as many types of defects as possible.

[0096] Specifically, speckle, as a relatively strong source of interference, is different from fog noise, and it is difficult to distinguish speckle from defects using a single channel.

[0097] In existing technologies, three separate detection paths (left, center, and right) are typically used. The left and right paths detect defects with large scattering angles, while the center path detects a type of defect in a specific direction. If only one of the left or right paths detects an imaging abnormality, it may be speckle interference. Only when both paths detect the defect simultaneously is it considered a defect.

[0098] The detection system 100 provided by this invention is backward compatible with traditional technology. For defects affected by speckle interference, the detection principle is as follows: if only one imaging module of the left imaging module 71 and the right imaging module 72 detects an imaging abnormality, then the abnormality is speckle interference. Only imaging abnormalities detected by both the left imaging module 71 and the right imaging module 72 are judged as defects.

[0099] Meanwhile, the detection system 100 provided by the present invention creatively sets at least two imaging channels 10 in an imaging module 1, that is, there are three channels in one detection. By adjusting the rotation angle of the light modulation component 12 in each imaging channel 10, the brightness of the defect is different in different imaging channels 10, while the effect on speckle is small. The difference between the images in the two imaging channels 10 can eliminate speckle and background noise, thereby obtaining an imaging image with a stronger defect signal.

[0100] Understandably, the detection system 100 is backward compatible with traditional technologies and can detect defects through different methods such as differential subtraction of the imaging signals of the two imaging channels 10 of an imaging module 1 or combination of the left imaging module 71 and the right imaging module 72. Furthermore, the three channels of the middle imaging module 6, the left imaging module 71 and the right imaging module 72 can work independently, and can be optimized for different types of small defects to improve detection efficiency and accuracy.

[0101] Furthermore, please refer to Figure 3 and Figure 5 , Figure 5 The detection system 100 shown includes a central imaging module 6 comprising imaging camera A1, imaging camera B1, and imaging camera C1; a left imaging module 71 comprising imaging camera A2, imaging camera B2, and imaging camera C2; and a right imaging module 72 comprising imaging camera A3, imaging camera B3, and imaging camera C3.

[0102] The detection system 100 provided by the present invention is based on different imaging channels 10, and can flexibly configure each imaging camera 11 to detect defects according to specific detection requirements. The detection system 100 of the present invention can meet at least two different detection configuration requirements.

[0103] Specifically, exemplarily, the first detection configuration is as follows: the central imaging module 6, the left imaging module 71, and the right imaging module 72 operate independently, and the imaging cameras A1, B1, and C1 in the central imaging module 6, A2, B2, and C2 in the left imaging module 71, and A3, B3, and C3 in the right imaging module 72 perform independent detection. Independent imaging is performed in each of the imaging cameras A1, B1, C1, A2, B2, C2, A3, B3, and C3, detecting at least nine types of defects and obtaining defect images. Each different imaging channel 10 performs polarization modulation and light intensity modulation on the light modulation component 12 set within the imaging channel 10 for its respective detected defect target to obtain better imaging of different types of defects.

[0104] Specifically, the second detection configuration is as follows: different imaging channels 10 of the same imaging module 1 detect different types of defects through differential imaging via the light modulation component 12.

[0105] By way of example, as a non-limiting specific implementation, in order to achieve the second detection configuration described above, the process of defect detection in the middle imaging module 6, left imaging module 71 and right imaging module 72 of the detection system 100 in this invention is shown as follows: The central imaging module 6, the left imaging module 71, and the right imaging module 72 each include three imaging channels 10, and the imaging cameras 11 in the corresponding imaging channels are imaging camera A1, imaging camera B1, imaging camera C1, imaging camera A2, imaging camera B2, imaging camera C2, imaging camera A3, imaging camera B3, and imaging camera C3, respectively.

[0106] The mid-path imaging module 6 is used to detect Type I and Type II small defects. After the illumination light is reflected by the mirror 14 of the illumination module 4 onto the surface of the wafer 8, the surface of the wafer 8 scatters the emitted light. The mid-path imaging module 6 collects the normally scattered light scattered at the normal angle. After the emitted light is initially focused by the imaging objective lens 13, it is split into three beams by the beam splitter 15 and enters the corresponding imaging channels 10. In their respective imaging channels 10, the polarization modulation component 121 corresponding to camera C1 is rotated to a certain angle to obtain that Type I and Type II small defects are relatively weak or disappear at the same time. The polarization modulation component 121 corresponding to camera A1 is rotated to a certain angle to obtain the strongest signal of Type I small defects. The polarization modulation component 121 corresponding to camera B1 is rotated to a certain angle to obtain the strongest signal of Type II small defects. The light intensity modulation component 122 in the three imaging channels 10 is adjusted so that the background noise intensity in the three imaging channels 10 is the same.

[0107] Furthermore, such as Figure 9 The schematic diagram shows the imaging results of imaging cameras A1, B1, and C1: the imaging of camera A1 includes defect I (signal visible) and background noise (haze noise, speckle); the imaging of camera B1 includes defect II (signal submerged in speckle, difficult to identify) and background noise (consistent with the background noise in camera A); the imaging of camera C1 is defect-free and contains only background noise consistent with that of cameras A1 and B1. By analyzing the imaging of camera C1, pure background noise features can be extracted. Using these features, differential calculations are performed on the images of cameras A1 and B1 respectively. Subtracting the imaging of camera C1 from the imaging of camera A1 yields defect I, and subtracting the imaging of camera C1 from the imaging of camera B1 yields defect II.

[0108] Furthermore, by using different imaging cameras A2, B2, C2, A3, B3, and C3 in different imaging channels 10 in the left imaging module 71 and right imaging module 72, defects of different types than those in the middle imaging module 6 can be obtained in the left imaging module 71 and right imaging module 72, respectively.

[0109] Specifically, the third detection configuration involves different imaging channels 10 of different imaging modules 1 detecting different types of defects through differential imaging by adjusting their respective corresponding light modulation components 12.

[0110] As a non-limiting specific implementation, in order to achieve the third detection configuration described above, the defect detection process in the middle imaging module 6, left imaging module 71, and right imaging module 72 of the detection system 100 provided in this embodiment is shown as follows: Two of the imaging channels 10 of the left imaging module 71 are used to detect small defect III. The outgoing light scattered from the surface of the wafer 8 is initially focused by the imaging objective lens 13 and then split into three beams by the beam splitter 15. Two of the beams carrying the image signal of small defect III are reflected by the mirror 14 and enter the imaging channels 10 corresponding to the imaging cameras A2 and B2, respectively. The two imaging channels 10 are differentially processed to eliminate speckle and background noise interference, thereby highlighting the characteristics of small defect III and obtaining type III defects. Two of the imaging channels 10 of the right imaging module 72 are used to detect small defects IV. The outgoing light scattered from the surface of the wafer 8 is initially focused by the imaging objective lens 13 and then split into three beams by the beam splitter 15. The two beams carrying the image signal of small defects IV are reflected by the mirror 14 and enter the imaging channels 10 corresponding to the imaging cameras A3 and B3, respectively. The two imaging channels 10 are differentially processed to eliminate speckle and background noise interference, thereby highlighting the characteristics of small defects IV and obtaining Class IV defects. Furthermore, the remaining imaging channel 10 in the left imaging module 71 and the remaining imaging channel 10 in the right imaging module 72 are used to detect small defect V. The left imaging module 71 and the right imaging module 72 respectively collect large-angle scattered light. After the illumination light is reflected by the reflector 14 of the illumination module 4 onto the surface of the wafer 8, the surface of the wafer 8 scatters the emitted light. The emitted light is initially focused by the imaging objective lens 13 and split into three beams by the beam splitter 15. Among them, the remaining beam of light carrying small defect V in the left imaging module 71 directly exits the beam splitter 15 and enters the imaging channel 10. After modulation, it is received by the imaging camera C2 and used to detect defect V. The remaining beam of light carrying small defect V in the right imaging module 72 directly exits the beam splitter 15 and enters the imaging channel 10. After modulation, it is received by the imaging camera C3. When the light signals collected by the imaging cameras C2 and C3 both contain the signal of defect V, it is determined that the imaging signal is a type V defect, rather than a speckle signal.

[0111] Two imaging channels 10 of the central imaging module 6 are used to detect small defect VI. After the normal scattered light is initially focused by the imaging objective lens 13, it is split into three beams by the beam splitter 15. Two of the beams carrying the image signal of small defect VI are reflected by the mirror 14 and enter the imaging channels 10 corresponding to the imaging cameras A1 and B1, respectively. They are then modulated by the polarization modulation component 121 and the light intensity modulation component 122 in sequence to obtain two sets of image signals for detecting small defect VI. One set of imaging signals has a stronger defect signal, and the other set has a weaker defect signal. The two channel images are differentially processed to eliminate speckle and background noise interference, thereby highlighting the characteristics of small defect VI and obtaining Class VI defects. The remaining imaging channel 10 in the central imaging module 6 is used to independently detect defect VII.

[0112] Furthermore, in this embodiment, the classification system of defects from Class I to Class VI is only for the purpose of explaining the technical implementation under different detection configurations. It does not limit the physical properties or optical characterization of various defects to be independent of each other. In actual application scenarios, based on the adjustment of the configuration parameters of the detection system 100 and the optimization of the detection strategy, the defect types targeted by different imaging channels 10 can overlap or be distinguished independently. The specific defect type is defined based on the actual detection requirements, wafer process characteristics and optical imaging analysis results, and the defect category attributes corresponding to each imaging channel 10 are flexibly determined.

[0113] Please see Figure 10 The second embodiment of the present invention provides a method for detecting surface defects on a wafer 8. This method is applied to the wafer 8 surface defect detection system 100 of the first embodiment, and the detection method includes the following steps: Step S1: Provide at least two imaging channels 10, acquire the light signal in the imaging channel 10, and convert the light signal into an image signal. The imaging channel 10 is provided with a light modulation component 12. Step S2: Adjust the angle of the light modulation component 12 in each imaging channel 10 independently using a preset angle group; Step S3: Obtain at least two sets of image signals corresponding to each imaging channel 10, perform differential operation on the at least two sets of image signals, and obtain the imaging image.

[0114] Understandably, the optical modulation component 12 is used to perform polarization modulation and intensity suppression on the outgoing light reflected or scattered by the surface of the wafer 8. Different defects, such as scratches, particles, dirt, bridging, short circuits, and subsurface stacking faults and dislocations, respond to the polarization and intensity of light. The outgoing light reflected or scattered by the surface of the wafer 8 contains different optical signals under different polarization parameters. The outgoing light in each imaging channel 10 forms different imaging images based on the rotation angle of different polarization modulation components 121. An imaging image with obvious contrast between defects and background noise and a strong defect signal is obtained in at least one imaging channel 10, and an imaging image with a weak defect signal is obtained in another imaging channel 10. After differential operation, the background noise is canceled or suppressed.

[0115] Specifically, in this invention, each step in the second embodiment can be executed by each component module in the first embodiment. Step S1 is executed by the imaging module 1, step S2 is executed by the polarization control module 2, and step S3 is executed by the signal processing module 3. The detection method used in this invention has the same beneficial effects as the detection system 100, and will not be described in detail here.

[0116] Furthermore, the imaging channel 10 includes a first imaging channel 101 and a second imaging channel 102, the preset angle group includes a first angle and a second angle, and the image signal includes a first image signal acquired based on the first imaging channel 101 and a second image signal acquired based on the second imaging channel 102.

[0117] Step S2, which independently adjusts the angle of the light modulation component 12 in each imaging channel 10 according to a preset angle group, includes the following steps: Step S21: Adjust the light modulation component 12 in the first imaging channel 101 to a first angle where the defect signal reaches its maximum value, and adjust the light modulation component 12 in the second imaging channel 102 to a second angle where the defect signal reaches its minimum value.

[0118] Understandably, the optical modulation components 12 in the first imaging channel 101 and the second imaging channel 102 are respectively set at a first angle and a second angle, with the aim of constructing a dual-channel image that maximizes the difference in defect signals, so as to provide data for subsequent differential operations.

[0119] Specifically, step S21 can be executed by the polarization control module 2 in the first embodiment, and has the same beneficial effects as the polarization control module 2 in the first embodiment, which will not be described again here.

[0120] Furthermore, step S2, which independently adjusts the angle of the light modulation component 12 in each imaging channel 10 according to a preset angle group, also includes the following steps: Step S22: Adjust the light intensity of the first imaging channel 101 and the second imaging channel 102 to make the background noise intensity of the first image signal and the second image signal consistent.

[0121] Understandably, in step S21, the difference between the defect signals acquired by the first imaging channel 101 and the second imaging channel 102 is maximized by adjusting the angle of the light modulation component 12. Maximizing the difference in defect signals improves the defect recognition capability after differential operation. In step S21, the light intensity of the first imaging channel 101 and the second imaging channel 102 is adjusted to make the background noise intensity of the first image signal and the second image signal consistent, thereby minimizing the residual background noise. Under the ideal state of consistent background noise, when differential operation is performed by the signal processing module 3, the background noise under the same light intensity control is canceled or weakened.

[0122] Specifically, background noise such as uniform speckle, haze noise, and film scattering has a relatively uniform polarization state and is less affected by differences in polarization parameters in different imaging channels 10. Their main difference comes from light intensity. The light intensity in the imaging channel 10 can be adjusted by means of light intensity regulation and power adjustment of the illumination source, so that the imaging brightness of background noise in different imaging channels 10 is consistent, and then the background noise is removed when the signal processing module 3 performs differential operation.

[0123] Further, in step S3, acquiring at least two sets of image signals corresponding to each imaging channel 10, performing differential operations on the at least two sets of image signals to acquire the imaging image also includes the following steps: Step S31: Analyze and calculate the signal-to-noise ratio of at least two sets of image signals to obtain the differential signal-to-noise ratio of different image signals.

[0124] Understandably, signal-to-noise ratio (SNR) is the ratio of target signal power to background noise power in an image signal. Differential SNR refers to the SNR of the difference between image signals in at least two imaging channels 10. The differential SNR can be set to a threshold. When the differential SNR obtained after being modulated by the optical modulation component 12 exceeds the threshold, it can be determined as a suitable setting angle, i.e., a preset angle group.

[0125] Specifically, step S31 can be processed by the signal processing module 3 in the first embodiment, which has the same beneficial effects as the polarization control module 2 in the first embodiment, and will not be described again here.

[0126] Furthermore, in this method, before step S1 is executed, there is also a related step of obtaining the rotation angle of the light modulation component 12 in the first imaging channel 101 and the second imaging channel 102, namely step S4.

[0127] Specifically, this application also includes step S4, which involves traversing the rotation angle combinations of the optical modulation components 12 in at least two imaging channels 10, calculating the differential signal-to-noise ratio under each angle combination, and obtaining angle combinations that are greater than a preset threshold.

[0128] Specifically, step S4 can be executed by a separate computer program or by the preprocessing component in signal processing module 3.

[0129] As a non-limiting embodiment, when executed by a separate computer program, the traversal of the rotation angle group is based on a Python console program. The range of rotation angle combinations and wafer type are specified by command-line parameters. The program outputs the optimal rotation angle combination configuration. The angle of the light adjustment component is pre-adjusted before the detection program starts. For the same batch of wafers 8 to be inspected, the angle of the light adjustment component is set only once for each program channel. When detecting one type of defect, the optimal rotation angle of the light adjustment component in one imaging channel 10 is used. When detecting another type of defect, the optimal rotation angle of the light adjustment component in another imaging channel 10 is used.

[0130] Understandably, the execution flow of step S4 is exemplarily shown in the first embodiment of the present invention and has the same beneficial effects as the traversal search execution flow in the first embodiment, and will not be described in detail here.

[0131] Furthermore, the second embodiment of the present invention provides a detection method in which the control of the polarization modulation component 121 and the light intensity modulation component 122 is used to achieve at least three different detection configuration requirements as in the first embodiment. In view of the differences in optical scattering characteristics of different types of defects, a multi-dimensional parameter control model is constructed to optimize the defect detection sensitivity, resolution and signal-to-noise ratio.

[0132] Understandably, the detection method provided by the present invention can switch the detection configuration in real time according to the material properties and defect distribution of different areas on the wafer surface, based on the experience of those skilled in the art or specific detection needs, to achieve multi-mode, multi-scale and multi-type defect detection on the same wafer.

[0133] Furthermore, the detection method described in this embodiment can be further combined with artificial intelligence image recognition algorithms. By performing feature extraction and machine learning training on the image data after polarization modulation and light intensity modulation, a mapping relationship between defect type and modulation parameters and detection configuration is established, thereby achieving adaptive optimization of the detection configuration. In subsequent iterations and upgrades, this detection method can be automatically extended to chip detection scenarios with different process technologies, compatible with various detection needs, and fully demonstrating the universality and scalability of the detection method provided by this invention.

[0134] Furthermore, there is no fixed execution order between steps S1 and S4; they can be executed in any order according to actual needs, with some steps being performed simultaneously.

[0135] Compared with the prior art, the wafer surface defect detection system and method provided by the present invention have the following beneficial effects: 1. The detection system provided in this embodiment of the invention uses an optical modulation component to modulate the polarization state and intensity of light reflected or scattered from the wafer surface, capturing the modulated polarized light to identify defects in the wafer. For smaller defects, the weak signal may mix with background noise. The imaging module sets at least two imaging channels, and a polarization control module sets different polarization control parameters and intensity parameters for the optical modulation component in each imaging channel. This allows multiple sets of image signals about the same area to be acquired simultaneously through different imaging channels, forming multiple polarized images and providing multi-dimensional data information for subsequent small defect extraction. The signal processing module performs differential operations on the image signals about the same area from different imaging channels, eliminating background noise such as speckle and haze, retaining only the difference signal caused by small defects. By combining the differential results of multiple image signals about the same area within multiple imaging channels, the features of small defects can be reconstructed, and other information about the small defects can be obtained. Therefore, the detection system provided by this invention, through the collaborative design of multiple imaging channels and multi-dimensional polarization modulation, is significantly superior to single-channel intensity imaging technology. Compared with the limitation of a single channel being able to detect only one type of small defect, the multiple imaging channels of this system can simultaneously perform parallel detection of multiple types of small defects. With the dynamic control of light polarization state and light intensity parameters by the polarization modulation component, the simultaneous acquisition and fusion calculation of multiple imaging channels improves the sensitivity and reliability of different types of defect detection. In terms of detection efficiency, the parallel operation of multiple imaging channels within the same imaging module changes the inefficient mechanism of single-channel multi-defect scanning, greatly shortening the detection time of multiple types of defects and adapting to the detection needs of different types of defects in wafer mass production scenarios.

[0136] 2. In this embodiment of the invention, the polarization control module dynamically and independently adjusts the light modulation components in the first and second imaging channels to a first angle and a second angle, thereby optimizing the polarization state of the light according to the characteristics of the defect. This adjusts the angle to the maximum or minimum value of the defect signal obtained by the signal processing module, significantly improving the contrast between the defect and the background. The polarization control module fixes the optimal polarization angle, reducing interference from ambient polarized light on the random polarization components of the detection system, making the detection results more stable and reducing the false detection rate. When the light modulation component is adjusted to the first angle where the defect signal is at its maximum value, the first imaging channel has the strongest response to the polarized light of the defect, resulting in a more accurate image of the defect. The brightness or grayscale value of the defect differs from the rest of the area. The defect appears brightest in the imaging image. When the light modulation component is adjusted to the second angle where the defect signal is at its minimum, the defect signal is suppressed to its minimum value. The imaging of the defect signal at this angle is weakened, and the defect is not displayed or has the weakest brightness in the imaging image. At this time, the imaging image of the second imaging channel can be used as a reference image to remove the interference of background noise. Therefore, by performing a differential operation with the imaging image of the first imaging channel, the common signals of background noise can be further eliminated, and only the characteristic differences of the defect are retained, thereby improving the sensitivity of defect detection. It is suitable for the detection of small defects in complex backgrounds.

[0137] 3. In this embodiment of the invention, speckle and haze noise are noises with no or weak polarization characteristics. Changes in the parameters of the optical modulation component have little impact on the imaging intensity of speckle and haze noise or are asynchronous with the small defect polarization imaging signal. For noises with polarization characteristics such as speckle and haze noise, the light intensity in different imaging channels is adjusted by different light transmittance of the optical attenuation element, so that the brightness of speckle and haze noise is consistent in different imaging channels, avoiding the impact of such polarization noise on detection accuracy. At least two sets of optical attenuation elements with different light transmittance continuously or gradiently adjust the light intensity in the imaging channel to adapt to the defect scattering characteristics and detection accuracy of different wafer surfaces. The through-hole without optical attenuation element has 100% light transmittance and is used when the strongest light signal is required, ensuring that the imaging module receives a sufficiently strong image signal and avoiding image blurring or information loss due to insufficient light intensity. The switching device is used to quickly switch the optical intensity modulation component to a no-attenuation state or different attenuation states, reducing the complexity of the detection system, reducing optical path offset error, and ensuring the stability and repeatability of optical intensity modulation.

[0138] 4. In this embodiment of the invention, the central imaging module collects reflected or scattered light in the direction perpendicular to the wafer surface, i.e., the normal direction, mainly detecting defects perpendicular or nearly perpendicular to the wafer surface, such as surface protrusions, depressions, and vertical cracks. These defects show significant changes in the intensity of scattered light in the normal direction. At the same time, the central imaging module provides a coordinate reference for the side imaging modules, facilitating the calibration of spatial positions for multiple imaging channels. The side imaging modules collect large-angle emitted light at a certain angle to the normal of the wafer surface, detecting defects with tilt angles, such as sidewall damage, oblique cracks, and edge foreign objects. The central and side imaging modules comprehensively detect defects from multiple angles. When only one imaging module works, such as the central module, the system can maintain basic defect detection. When two or three imaging modules work simultaneously, they work together to improve detection accuracy through multi-source data fusion.

[0139] 5. In this embodiment of the invention, during wafer inspection, the illumination module controls the polarization state of the illumination light to be at least one of parallel polarized light, vertical polarized light, or elliptically polarized light, thereby adjusting the interaction between the illumination light and the wafer surface structure, thus optimizing the contrast, sensitivity, and defect identification capability of the detection signal. For example, when parallel polarized light and vertical polarized light are used for simultaneous illumination, data signals of surface flatness and defect contrast can be acquired simultaneously, improving the accuracy of defect classification.

[0140] 6. This invention also provides a detection method applied to the above-mentioned detection system. In this method, the angle of the light modulation component is changed in at least two imaging channels. The polarization characteristics of background noise, such as speckle and haze noise, are usually consistent. However, different defects, such as scratches, particles, dirt, bridging, short circuits, and subsurface stacking faults and dislocations, have a greater polarization response to light than a light intensity response. The light emitted after reflection or scattering from the wafer surface contains different light signals under different polarization parameters. The emitted light in each imaging channel forms different imaging images based on different polarization angles. In at least one imaging channel, an image with a significant contrast between the defect and background noise and a strong defect signal is obtained. In another imaging channel, an image with a weak defect signal is obtained. After differential operation, the background noise is canceled out, and the image signal of the defect is displayed. In this method, polarization differential imaging of at least two imaging channels breaks through the limitations of detection with a single imaging channel. It uses the difference in polarization response rather than the difference in light intensity to identify defects and suppresses common background noise in the same area through differential operation, adapting to complex detection environments.

[0141] 7. In this embodiment of the invention, a single imaging channel can only acquire signals at a fixed polarization angle. Defect identification relies on the intensity or phase change of light under a single polarization state, which is easily affected by background noise and cannot distinguish between defects and related noise. However, the multiple imaging channels in this invention focus on the polarization change caused by the defect and suppress the background polarized light, highlighting the polarization difference between the defect and the background noise. The first imaging channel and the second imaging channel obtain the signal difference between the maximum and minimum values ​​of the defect signal through the angle of the modulation light modulation component. After differential operation by the signal processor, the background noise is not obvious or is not displayed at all, and the defect signal can be separated from the background noise.

[0142] 8. In this embodiment of the invention, the difference in adjustment of defects and background noise by the optical modulation components in different imaging channels is utilized. The intensity of background noise is made consistent in the two imaging channels by adjusting the optical intensity. Then, the background noise is canceled out by signal subtraction, and the defect signal is retained. The background noise includes speckle and haze noise. Speckle is random bright and dark spots generated by laser interference. Haze noise is diffuse scattering background caused by small undulations or impurities on the wafer surface. When the intensity of the two imaging channels is adjusted to make the background noise consistent, the speckle and haze noise will cancel each other out in the differential operation due to the consistent distribution in different channels of the same imaging module.

[0143] 9. In this embodiment of the invention, the signal-to-noise ratio (SNR) is the ratio of the target signal power to the background noise power in the image signal. The differential SNR refers to the difference in SNR between the differential signals of the images in at least two imaging channels. The differential SNR can be set to a threshold. When the differential SNR obtained after the optical modulation component rotates by an angle exceeds the threshold, it can be determined as a better setting angle, i.e., a preset angle group. In this invention, the differential SNR is upgraded from the single SNR of a single imaging channel to the differential SNR difference assessment between multiple imaging channels through comparative analysis. The ability to identify defects is improved by the differential SNR difference characteristics, and a quantitative comparison basis is provided for multiple sets of data obtained from multiple imaging channels.

[0144] 10. In this embodiment of the invention, by traversing all possible combinations of rotation angles of the optical modulation components, the optimal setting angle that makes the differential signal-to-noise ratio exceed a threshold is found. For example, when the angle of the optical modulation component in the first imaging channel is A, the angle of the optical modulation component in the second imaging channel is rotated 360° and the differential signal-to-noise ratio with that of the image signal in the first imaging channel is calculated. When the angle of the optical modulation component in the second imaging channel is B, and the differential signal-to-noise ratio reaches or exceeds the preset differential signal-to-noise ratio threshold, the defect signals of the image signals in the first and second imaging channels are significantly amplified through differential operation, while background noise such as speckle and fog noise is suppressed. The above provides a detailed description of a wafer surface defect detection system and method disclosed in the embodiments of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention. Any modifications, equivalent substitutions, and improvements made within 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, characterized in that: include: An imaging module, comprising at least two imaging channels, each imaging channel including an imaging camera that acquires optical signals and converts the optical signals into image signals, and an optical modulation component disposed within the imaging channel, the optical modulation component being used to perform polarization modulation and intensity modulation on the outgoing light reflected or scattered by the wafer surface. A polarization control module is used to independently adjust the rotation angle of the light modulation component in each imaging channel according to a preset angle group. The signal processing module is connected to the imaging module. The signal processing module is used to acquire the image signal corresponding to each imaging channel and perform differential operation on the different image signals to acquire the imaging image.

2. The wafer surface defect detection system as described in claim 1, characterized in that: The imaging channel includes a first imaging channel and a second imaging channel, the preset angle group includes a first angle and a second angle, the image signal includes a defect signal, the optical modulation component includes a polarization modulation component, and the polarization control module is used to adjust the polarization modulation component in the first imaging channel to the first angle where the defect signal reaches its maximum value, and to adjust the polarization modulation component in the second imaging channel to the second angle where the defect signal reaches its minimum value.

3. The wafer surface defect detection system as described in claim 1, characterized in that: The optical modulation component includes an optical intensity modulation component, which includes at least two sets of optical attenuation elements with different attenuation rates and at least one light-transmitting channel without the optical attenuation elements. The imaging module includes a switching device, and the optical intensity modulation component is mounted on the switching device.

4. The wafer surface defect detection system as described in claim 1, characterized in that: The detection system includes at least one side-channel imaging module and a center-channel imaging module. The direction perpendicular to the wafer surface is defined as the wafer normal. The center-channel imaging module collects the outgoing light reflected or scattered from the wafer normal, and the side-channel imaging module collects the outgoing light that forms an angle with the wafer normal.

5. The wafer surface defect detection system as described in claim 1, characterized in that: The system includes an illumination module for providing illumination light to illuminate the wafer to be inspected, the illumination light including at least one of parallel polarized light, vertical polarized light, or elliptically polarized light.

6. A method for detecting defects on a wafer surface, characterized in that, The wafer surface defect detection system as described in any one of claims 1-5, the wafer surface defect detection method comprising the following steps: It provides at least two imaging channels, acquires optical signals within the imaging channels, and converts the optical signals into image signals. An optical modulation component is provided within the imaging channels. The rotation angle of the light modulation component in each imaging channel can be independently adjusted using a preset angle group; At least two sets of image signals corresponding to each of the imaging channels are acquired, and differential operations are performed on the at least two sets of image signals to acquire an imaging image.

7. The method for detecting wafer surface defects as described in claim 6, characterized in that: The imaging channel includes a first imaging channel and a second imaging channel, the preset angle group includes a first angle and a second angle, the image signal includes a defect signal, and adjusting the rotation angle of the optical modulation component in each imaging channel independently according to the preset angle group includes the following steps: Adjust the light modulation component in the first imaging channel to the first angle at which the defect signal reaches its maximum value, and adjust the light modulation component in the second imaging channel to the second angle at which the defect signal reaches its minimum value.

8. The method for detecting wafer surface defects as described in claim 7, characterized in that: The image signal includes a first image signal acquired based on the first imaging channel and a second image signal acquired based on the second imaging channel. The step of independently adjusting the rotation angle of the light modulation component within each imaging channel according to a preset angle group further includes the following steps: Adjust the light intensity of the first imaging channel and the second imaging channel to make the background noise intensity of the first image signal and the second image signal consistent.

9. The method for detecting wafer surface defects as described in claim 6, characterized in that: The process of acquiring at least two sets of image signals corresponding to each of the imaging channels, performing a difference operation on the at least two sets of image signals, and acquiring an imaging image further includes the following steps: The signal-to-noise ratios of at least two sets of the image signals are analyzed and calculated to obtain the differential signal-to-noise ratios of the different image signals.

10. The method for detecting wafer surface defects as described in claim 9, characterized in that: The method further includes the following steps: traversing at least two sets of rotation angle combinations of the optical modulation components in the imaging channels, calculating the differential signal-to-noise ratio under each angle combination, and obtaining angle combinations greater than a preset threshold.

Images