Silicon through hole defect detection method and system

Abstract

The invention provides a silicon through hole defect detection method and system, and relates to the field of silicon through hole defect detection. According to the method, the light spot image formed on the optical receiving part after the through-silicon-via wafer is irradiated by the light source is firstly obtained, then the light spot feature of the light spot image is extracted, the definition index is generated, and then when the definition index is lower than the target index threshold value, the parameters of the light source and / or the wafer bearing part used for bearing the through-silicon-via wafer are adjusted according to the definition index. And finally, when the definition index is equal to or higher than the target index threshold, performing defect detection on the through silicon via of the through silicon via wafer. According to the method, misjudgment or missing detection caused by image blurring can be avoided, the accuracy of defect detection is improved, point-by-point or sampling detection is not needed, rapid full-coverage detection of the through silicon vias of the whole wafer is achieved, therefore, the detection efficiency is remarkably improved, and the process requirement of large-scale mass production is met.

Description

Technical Field

[0001] This invention relates to the field of through-silicon via (TSV) defect detection technology, and in particular to a method and system for detecting TSV defects. Background Technology

[0002] In semiconductor manufacturing, through-silicon vias (TSVs) are a key process for achieving three-dimensional integrated circuit interconnects, and their processing quality directly affects the electrical performance and reliability of the chip. Existing TSV defect detection methods mostly employ point-by-point scanning or local imaging to determine defects in the vias on the wafer. However, these methods often suffer from low detection efficiency and limited coverage, making it difficult to comprehensively reflect the overall quality of the entire wafer in a timely manner. Furthermore, due to the tiny diameter and large depth-to-width ratio of TSVs, existing optical inspection methods that rely solely on the number of light spots to determine wafer defects are susceptible to interference from unclear light spot images, leading to misjudgments.

[0003] Therefore, existing methods for detecting defects in through-silicon vias (TSVs) suffer from low accuracy and inefficiency when inspecting TSV wafers. Summary of the Invention

[0004] One objective of this invention is to provide a method for detecting through-silicon via (TSV) defects, thereby addressing the technical problems of low accuracy and low efficiency in existing TSV defect detection methods.

[0005] Another object of the present invention is to provide a through-silicon via (TSV) defect detection system.

[0006] Specifically, the present invention provides a method for detecting defects in through-silicon vias (TSVs), comprising the following steps: Acquire an image of the light spot formed on the optical receiver after a through-silicon via wafer is illuminated by a light source; Extract the spot features from the spot image to generate a sharpness index; When the sharpness index is lower than the target index threshold, the parameters of the light source and / or the wafer carrier used to support the through-silicon via wafer are adjusted according to the sharpness index. When the clarity index is equal to or higher than the target index threshold, defect detection is performed on the through-silicon via (TSV) of the TSV wafer.

[0007] Optionally, the step of extracting the spot features of the spot image and generating a sharpness index includes the following steps: Extract the edge sharpness feature S of the light spot in the light spot image. edge Central contrast feature C contrast Shape roundness feature R round and internal grayscale uniformity characteristics Uuniform ; The sharpness index, used to characterize the sharpness of a light spot, is calculated using the following formula. C I =w1×S edge +w2×C contrast +w3×R round +w4×U uniform ; Where w1 is the weight coefficient of edge sharpness feature, w2 is the weight coefficient of center contrast feature, w3 is the weight coefficient of shape roundness feature, and w4 is the weight coefficient of internal gray level uniformity feature.

[0008] Optionally, the method for dynamically adjusting the weight coefficients of each of the light spot features includes the following steps: Obtain the initial weighting coefficients for each of the aforementioned light spot features; Based on real-time analysis of the correlation between multiple spot features and defect determination results, the contribution of each spot feature to the defect determination results is determined. The initial weight coefficients of the corresponding spot features are corrected based on the contribution level to obtain updated weight coefficients.

[0009] Optionally, in the step of adjusting the parameters of the light source and / or the wafer carrier for supporting the through-silicon via wafer according to the sharpness index when the sharpness index is lower than the target index threshold, the parameters of the light source include the light source brightness, position and illumination angle, and the parameters of the wafer carrier include the height of the wafer carrier.

[0010] Optionally, the parameters of the light source and the wafer carrier are adjusted as follows: When the brightness of the light spot in the light spot image is insufficient or oversaturated, the brightness of the light source is adjusted; When the edge of the light spot becomes blurred, adjust the distance between the light source and the through-silicon via wafer; When the shape of the light spot shifts or becomes distorted, the illumination angle of the light source is adjusted. When the light spot image is out of focus, adjust the height of the wafer carrier.

[0011] Optionally, when the sharpness index is equal to or higher than the target index threshold, the step of performing defect detection on the through-silicon via (TSV) of the TSV wafer includes the following steps: The number of detected light spots is compared with the actual number of through-silicon vias to determine whether there are any non-through defects in the through-silicon vias; The diameter and / or brightness of the light spot in the light spot image are measured and compared with a preset threshold range to determine whether the through-silicon via has a blockage defect or a partial non-contact defect.

[0012] In particular, the present invention also provides a through-silicon via (TSV) defect detection system, including a TSV defect detection device and a spot detection system for performing the TSV defect detection method as described above, wherein the TSV defect detection device includes: light source; Wafer carrier section, used to place through-silicon via (TSV) wafers; An optical receiving unit is used to receive the light spot formed after the light source illuminates the through-silicon via wafer.

[0013] Optionally, the light source is configured to be operatively or controllably moved along a vertical direction perpendicular to the wafer carrier. The wafer carrier is configured to be operably or controllably movable along the vertical direction.

[0014] Optionally, the wafer carrier has a clamping assembly configured to move laterally along a direction perpendicular to the vertical direction for clamping the through-silicon via wafer.

[0015] Optionally, the spot detection system includes: Image acquisition module, used to acquire a light spot image containing the light spot; The detection module, connected to the image acquisition module, is used to detect the spot image to determine whether the through-silicon via (TSV) of the TSV wafer has defects.

[0016] This invention first acquires an image of the light spot formed on the optical receiving section after the through-silicon via (TSV) wafer is illuminated by a light source. Then, the light spot features are extracted to generate a sharpness index. Subsequently, when the sharpness index is lower than a target threshold, the parameters of the light source and / or the wafer support section used to support the TSV wafer are adjusted according to the sharpness index. Finally, when the sharpness index is equal to or higher than the target threshold, defect detection is performed on the TSV wafer. This invention avoids misjudgments or missed detections caused by image blurring, improving the accuracy of defect detection. Furthermore, it eliminates the need for point-by-point or sampling inspection, achieving rapid, full-coverage inspection of all TSVs on the entire wafer, thus significantly improving inspection efficiency and meeting the process requirements of large-scale mass production.

[0017] The above and other objects, advantages and features of the present invention will become more apparent to those skilled in the art from the following detailed description of specific embodiments of the invention in conjunction with the accompanying drawings. Attached Figure Description

[0018] The following sections will describe some specific embodiments of the invention in detail by way of example and not limitation, with reference to the accompanying drawings. The same reference numerals in the drawings denote the same or similar parts or portions. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings: Figure 1 This is a schematic flowchart of a through-silicon via defect detection method according to an embodiment of the present invention; Figure 2 This is a schematic flowchart of a method for obtaining a sharpness index according to an embodiment of the present invention; Figure 3 This is a schematic flowchart of a method for adjusting the parameters of a light source and a wafer carrier according to an embodiment of the present invention; Figure 4 This is a schematic flowchart of a through-silicon via defect detection method according to an embodiment of the present invention; Figure 5 This is a schematic structural diagram of a silicon through-hole defect detection device according to an embodiment of the present invention; Figure 6 This is a schematic structural diagram of a spot detection system according to an embodiment of the present invention.

[0019] Figure label: 100-Through Silicon Via (TSV) Defect Detection Device, 200-TSV Wafer, 1-Light Source, 2-Wafer Carrier, 3-Optical Receiving Unit, 4-First Sliding Assembly, 5-Second Sliding Assembly, 6-Clamping Assembly, 61-Third Driving Component, 62-Limiting Plate. Detailed Implementation

[0020] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0021] Figure 1 This is a schematic flowchart of a through-silicon via (TSV) defect detection method according to an embodiment of the present invention. Figure 2 This is a schematic flowchart of a method for obtaining sharpness indicators according to an embodiment of the present invention. Figure 3 This is a schematic flowchart of a parameter adjustment method for a light source and a wafer carrier according to an embodiment of the present invention. Figure 4 This is a schematic flowchart of a through-silicon via (TSV) defect detection method according to an embodiment of the present invention. Figure 5 This is a schematic structural diagram of a through-silicon via (TSV) defect detection device according to an embodiment of the present invention. Figure 6 This is a schematic structural diagram of a spot detection system according to an embodiment of the present invention.

[0022] like Figure 1 As shown, in one specific embodiment, a method for detecting through-silicon via (TSV) defects includes the following steps: Step S100: Obtain the light spot image formed on the optical receiving unit 3 after the through silicon via wafer 200 is illuminated by the light source 1; Step S200: Extract the spot features of the spot image and generate a sharpness index; Step S300: When the sharpness index is lower than the target index threshold, adjust the parameters of the light source 1 and / or the wafer carrier 2 used to carry the through silicon via wafer 200 according to the sharpness index. Step S400: When the clarity index is equal to or higher than the target index threshold, perform defect detection on the through-silicon via (TSV) of the through-silicon via wafer 200.

[0023] In this embodiment of the invention, a light spot image formed on the optical receiving unit 3 after the through-silicon via (TSV) wafer 200 is illuminated by the light source 1 is first acquired. Then, the light spot features of the light spot image are extracted to generate a sharpness index. Subsequently, when the sharpness index is lower than the target index threshold, the parameters of the light source 1 and / or the wafer support unit 2 used to support the TSV wafer 200 are adjusted according to the sharpness index. Finally, when the sharpness index is equal to or higher than the target index threshold, defect detection is performed on the TSVs of the TSV wafer 200. The method of this embodiment of the invention can avoid misjudgment or missed detection caused by image blurring, improve the accuracy of defect detection, and achieve rapid full-coverage detection of TSVs on the entire wafer without point-by-point or sampling inspection, thereby significantly improving detection efficiency and meeting the process requirements of large-scale mass production.

[0024] In step S100, the light source 1 illuminates the upper surface of the through-silicon via (TSV) wafer 200 and forms a light spot on the optical receiving unit 3. The TSV diameter of the TSV wafer 200 is greater than 80 μm, and the operating wavelength of the light source 1 is 0.4 μm-1.5 μm, for example, 0.4 μm, 0.8 μm, 1 μm, or 1.5 μm. If the TSV diameter of the TSV wafer 200 is less than 80 μm, the diffraction phenomenon of the transmitted light is aggravated, resulting in severe dispersion of the light spot and blurred boundaries, making it impossible to form a clear light spot image, thus making the TSV defect detection results of the present invention inaccurate. However, the TSV diameter of the TSV wafer 200 is greater than 80 μm, which effectively avoids the diffraction effect of the transmitted light, so as to form a light spot with clear boundaries and concentrated energy on the optical receiving unit 3.

[0025] In step S200, the spot features of the spot image are extracted to generate a sharpness index, thereby transforming the spot features into a quantifiable mathematical description and establishing a bridge from optical imaging to data analysis. The sharpness index is strongly correlated with the actual physical state of the through-silicon via (TSV), providing an objective basis for determining whether there are defects in the TSV of the TSV wafer 200.

[0026] like Figure 2 As shown, in one embodiment, step S200 includes: Step S210: Extract the edge sharpness feature S of the light spot in the light spot image. edge Central contrast feature C contrast Shape roundness feature R round and internal grayscale uniformity characteristics U uniform ; Step S220: Calculate the sharpness index used to characterize the sharpness of the light spot according to the following formula. C I =w1×S edge +w2×C contrast +w3×R round +w4×U uniform ; Where w1 is the weight coefficient of edge sharpness feature, w2 is the weight coefficient of center contrast feature, w3 is the weight coefficient of shape roundness feature, and w4 is the weight coefficient of internal gray level uniformity feature.

[0027] In step S210, the edge sharpness feature is used to measure the transition from bright to dark at the edge of the light spot. The steeper the transition, the clearer the edge of the light spot, which can reflect whether the boundary of the light spot is complete. The center contrast feature characterizes the contrast of the light spot by comparing the brightness difference between the center and the surrounding area. The more obvious the difference, the clearer the light spot and the higher the signal-to-noise ratio. The shape roundness feature is used to evaluate how close the outline of the light spot is to an ideal circle. The higher the roundness, the more regular the formation of the through-silicon via. If there is a deviation, it reflects defects or abnormalities in the through-silicon via. The internal grayscale uniformity feature detects the consistency of the grayscale distribution inside the light spot and reflects whether there are local dark spots or bright spots inside the light spot. If the uniformity is high, it indicates that the transmitted light distribution is stable.

[0028] In step S220, edge sharpness features, center contrast features, shape roundness features, and internal grayscale uniformity features are combined into a single sharpness index by weighted calculation. This avoids the limitations of judging a single spot feature and ensures that the sharpness index comprehensively reflects the image quality of the spot.

[0029] In one embodiment, the method for dynamically adjusting the weight coefficients of each spot feature includes the following steps: Step S211: Obtain the initial weight coefficients for each spot feature; Step S212: Based on the real-time analysis of the correlation between multiple spot features and defect judgment results, determine the contribution of each spot feature to the defect judgment results. Step S213: Correct the initial weight coefficients of the corresponding spot features according to the contribution to obtain the updated weight coefficients.

[0030] In step S211, the initial weight coefficients of each spot feature are obtained. This typically refers to setting reference weight coefficient values ​​for the spot features based on historical experimental data, historical sample detection results, or expert experience during the initial setup or algorithm training phase. For example, by performing spot imaging analysis on a large number of through-silicon via (TSV) wafer 200 samples in known states, the discrimination capabilities of edge sharpness features, center contrast features, shape roundness features, and internal grayscale uniformity features in distinguishing between normal and defective TSVs are statistically analyzed, thereby determining the initial weight coefficients for edge sharpness features, center contrast features, shape roundness features, and internal grayscale uniformity features.

[0031] In step S212, the real-time change trends of edge sharpness features, center contrast features, shape roundness features, and internal grayscale uniformity features are collected and statistically analyzed in conjunction with the judgment results of the current through-silicon via wafer 200 inspection samples. The judgment result is either normal or defective. By calculating the correlation index between each spot feature and the defect judgment result, such as Pearson correlation coefficient, mutual information, feature sensitivity, or feature importance score based on gradient boosting model, the influence of each spot feature on the defect judgment result in the current detection scenario is quantified. The higher the correlation, the more significant the influence of the spot feature on the defect judgment result, and the higher its contribution is considered, thus guiding the subsequent dynamic correction of the initial weight coefficient.

[0032] In step S213, based on the contribution of each spot feature determined in step S212, the corresponding initial weight coefficients are corrected. Specifically, if a spot feature shows a high correlation in defect judgment, its corresponding weight coefficient is increased to enhance its influence in the sharpness index calculation; conversely, if a spot feature contributes little to the judgment result, its weight coefficient is appropriately decreased to reduce its interference. For example, if the change in edge sharpness features is highly consistent with the occurrence of silicon via defects in a batch of through-silicon via (TSV) wafers 200, it indicates that the feature contributes significantly and its corresponding weight coefficient should be increased; conversely, if a spot feature has a low correlation with the defect judgment result, its weight coefficient can be appropriately decreased. Through the above correction process, updated weight coefficients are obtained, enabling the calculated sharpness index to more accurately reflect the spot image quality.

[0033] For example, the initial weighting coefficients for edge sharpness, center contrast, shape roundness, and internal grayscale uniformity were 0.30, 0.25, 0.25, and 0.20, respectively. Subsequently, during the inspection of through-silicon via (TSV) defects in a batch of 200 TSV wafers, real-time analysis was performed on the statistical correlation between the fluctuation of each spot feature and the defect determination results. The correlation indices for edge sharpness, shape roundness, center contrast, and internal grayscale uniformity were found to be 0.88, 0.63, 0.42, and 0.27, respectively. Based on this, the edge sharpness feature was determined... The feature with the highest contribution is the internal grayscale uniformity feature, while the feature with the lowest contribution is the internal grayscale uniformity feature. The initial weight coefficients are then adjusted based on their contribution values. The weight coefficients for edge sharpness feature are adjusted from 0.30 to 0.47, for shape roundness feature from 0.25 to 0.31, for center contrast feature from 0.25 to 0.16, and for internal grayscale uniformity feature from 0.20 to 0.06. This results in updated weight coefficients, which make subsequent defect judgment processes focus on the spot features with higher contribution values, thus improving the reliability and accuracy of through-silicon via (TSV) defect judgment.

[0034] In step S300, when the clarity index is lower than the target index threshold, it indicates that the quality of the currently acquired spot image is insufficient and cannot meet the requirements of subsequent through-silicon via (TSV) defect detection. Based on the feedback result of the clarity index, the detection conditions are adaptively adjusted. On the one hand, the parameters of the light source 1 can be adjusted to enhance the imaging effect of the transmitted light. On the other hand, the parameters of the wafer carrier 2 carrying the TSV wafer 200 can be adjusted to optimize the vertical distance between the TSV wafer 200 and the optical receiving unit 3. Through the above dynamic adjustment, the clarity index of the spot image gradually meets the standard, thereby ensuring that the spot image that finally enters the TSV defect detection stage has reliability and stability.

[0035] In one embodiment, the parameters of the light source 1 include the light source brightness, position, and illumination angle, and the parameters of the wafer carrier 2 include the height of the wafer carrier 2. In this embodiment, when the sharpness index is lower than the target index threshold, the brightness, position, and illumination angle of the light source 1, as well as the height of the wafer carrier 2, are adjusted according to the sharpness index to maintain the quality of the light spot image at its best, thereby improving the accuracy of the defect determination result.

[0036] like Figure 3 As shown, in one embodiment, in step S300, the parameters of the light source 1 and the wafer carrier 2 are adjusted as follows: Step S310: When the brightness of the light spot in the light spot image is insufficient or oversaturated, adjust the brightness of the light source 1; Step S320: When the edge of the light spot becomes blurred, adjust the distance between the light source 1 and the through-silicon via wafer 200; Step S330: When the shape of the light spot shifts or becomes distorted, adjust the illumination angle of the light source 1; In step S340, when the spot image is out of focus, the height of the wafer carrier 2 is adjusted.

[0037] In step S310, when the brightness of the light spot image is detected to be insufficient or oversaturated, the brightness of the light source 1 is adjusted. Specifically, when the light spot brightness is insufficient, it indicates that the transmitted light signal is too weak, which may cause the edge of the light spot to be blurred or difficult to distinguish. At this time, by increasing the brightness of the light source 1, the intensity of transmitted light can be enhanced, thereby improving the clarity of the light spot. When the light spot brightness is oversaturated, the central area of ​​the light spot may experience brightness overflow, loss of detail, or decreased image contrast. At this time, by reducing the brightness of the light source 1, overexposure can be avoided, and the layers and details of the light spot image can be maintained. Through this dynamic brightness adjustment, the light source 1 is always kept within a suitable brightness range.

[0038] In step S320, when the edge of the light spot is blurred, the distance between the light source 1 and the through-silicon via (TSV) wafer 200 is adjusted so that the transmitted light can form a light spot on the optical receiver 3 at a more suitable divergence angle after passing through the TSV, thereby improving the focusing effect of the light spot. If the distance between the light source 1 and the TSV wafer 200 is too close, it may lead to uneven distribution of transmitted light and irregular boundaries of the formed light spot; if the distance is too far, it may cause excessive diffusion of the light spot and blurry edges. By dynamically adjusting the distance between the light source 1 and the TSV wafer 200, the edge of the light spot becomes clearer and sharper, ensuring that the extracted light spot features are accurate and reliable.

[0039] In step S330, when the shape of the light spot is shifted or distorted, it indicates that the path of the transmitted light and the axis of the through-silicon via (TSV) are not ideally aligned. By adjusting the illumination angle of the light source 1, the transmitted light passes through the TSV along a direction perpendicular to the TSV wafer 200, thereby ensuring that the distribution of the transmitted light on the optical receiving part 3 is consistent with the actual shape of the TSV.

[0040] In step S340, when the spot image is out of focus, it indicates that the position between the through-silicon via wafer 200 and the optical receiver 3 is not within the optimal imaging focal length range. By adjusting the height of the wafer carrier 2, the distance between the through-silicon via wafer 200 and the optical receiver 3 is changed, thereby obtaining a clear and stable spot image.

[0041] In step S400, when the clarity index is equal to or higher than the target index threshold, defect detection is performed on the through-silicon via wafer 200 to ensure that all through-holes on the entire through-silicon via wafer 200 are detected simultaneously, effectively improving the accuracy and stability of through-silicon via defect identification.

[0042] like Figure 4As shown, in one embodiment, step S400 includes: Step S410: Compare the number of detected light spots with the actual number of through-silicon vias to determine whether there are any non-through defects in the through-silicon vias; Step S420: Measure the diameter and / or brightness of the light spot in the light spot image and compare it with a preset threshold range to determine whether there are blockage defects or partial blockage defects in the through silicon via.

[0043] In step S410, the presence of non-through defects in the through-silicon via (TSV) wafer 200 is determined by comparing the number of light spots with the actual number of TSVs. Specifically, the TSV wafer 200 has a fixed and known total number of vias during its design. The number of light spots in the light spot image that meets the clarity index is obtained. If all TSVs are through, the number of light spots on the optical receiving unit 3 is consistent with the actual number of TSVs. If the number of detected light spots is less than the actual number of TSVs, it indicates that some TSVs have failed to transmit light, possibly due to incomplete processing, blockage, or residue covering. By comparing the number of detected light spots with the actual number of TSVs, TSVs with non-through defects can be quickly identified, avoiding reliance on point-by-point scanning detection and improving detection efficiency.

[0044] In step S420, the diameter and / or brightness of each spot in the spot image are acquired and compared with a preset threshold range to determine whether the corresponding through-silicon via (TSV) has blockage or partial blockage defects. Specifically, when a TSV has impurities, uneven metal filling, or deposits on its wall, its transmitted light will be weakened or blocked, resulting in a spot brightness lower than the preset threshold range. If the opening of the TSV is restricted or the aperture is reduced, the transmitted light will shrink, and the corresponding spot diameter will be significantly smaller than the preset threshold range. By incorporating the diameter and / or brightness of the spot into the detection criteria, the presence of non-through-hole defects in the through-hole transmission can be more comprehensively reflected. Complete blockage defects can be identified, as well as partial blockage or semi-blockage latent defects, thereby improving the accuracy of the detection.

[0045] like Figure 5As shown, in one embodiment, a through-silicon via (TSV) defect detection system includes a TSV defect detection device 100 and a spot detection system for performing the TSV defect detection method as described above. The TSV defect detection device 100 includes a light source 1, a wafer carrier 2, and an optical receiver 3. The wafer carrier 2 is used to place the TSV wafer 200, and the optical receiver 3 is used to receive the spot formed after the TSV wafer 200 is irradiated by the light source 1. In this embodiment, the optical receiver 3 is disposed below the TSV wafer 200, and the light source 1 is disposed above the TSV wafer 200. The light emitted by the light source 1 can be uniformly projected onto the entire upper surface of the TSV wafer 200, thereby forming a spot on the optical receiver 3. If the emitting surface of the light source 1 is circular, the diameter of the emitting surface of the light source 1 is 400 mm. If the emitting surface of the light source 1 is rectangular, the size of the emitting surface of the light source 1 is 400 mm × 400 mm, ensuring that the light field irradiated by the light source 1 onto the TSV wafer 200 is uniform and sufficient.

[0046] In one embodiment, the light source 1 is configured to be operably or controllably moved along a vertical direction perpendicular to the wafer carrier 2, and the wafer carrier 2 is configured to be operably or controllably moved along a vertical direction. The through-silicon via (TSV) defect detection device 100 further includes a first sliding assembly 4 and two sets of second sliding assemblies 5. The first sliding assembly 4 is slidably connected to the light source 1, thereby causing the light source 1 to move vertically. Each second sliding assembly 5 is slidably connected to one end of the wafer carrier 2, thereby causing the TSV wafer 200 to move vertically with the wafer carrier 2. Specifically, the first sliding assembly 4 includes a first sliding guide rail, a first driving member, and a slide rod. The first sliding guide rail is arranged vertically and slidably connected to the slide rod. The slide rod is connected to the first driving member and is configured to move vertically under the drive of the first driving member, thereby causing the light source 1 to move vertically, or allowing the operator to move the light source 1 in a vertical direction perpendicular to the wafer carrier 2. Each second sliding component 5 includes a second sliding guide rail, a second driving member, and a slide base. The second sliding guide rail is arranged vertically and slidably connected to the slide base. The slide base is connected to the second driving member and the wafer carrier 2 respectively, and is configured to move vertically under the drive of the second driving member, thereby moving the wafer carrier 2 vertically, or allowing the operator to move the wafer carrier 2 in the vertical direction.

[0047] In one embodiment, the wafer carrier 2 has a clamping assembly 6, which is configured to move laterally in a direction perpendicular to the vertical direction to clamp the through-silicon via (TSV) wafer 200. Specifically, the TSV defect detection device 100 further includes the clamping assembly 6 for clamping the TSV wafer 200. Specifically, the clamping assembly 6 includes a third driving member 61 and a limiting plate 62 arranged at intervals. The third driving member 61 is configured to apply a controlled lateral external force in a direction perpendicular to the vertical direction to the TSV wafer 200, so that the TSV wafer 200 abuts against the limiting plate 62. The wafer carrier 2 has a hollow structure extending along its thickness direction, on which the TSV wafer 200 is placed to ensure that the TSV wafer 200 does not fall off. The distance between the limiting plate 62 and the third driving member 61 is in the range of 100mm-400mm, for example, 100mm, 200mm, 300mm or 400mm.

[0048] like Figure 6 As shown, in one embodiment, the spot detection system includes an image acquisition module 8 and a detection module 9. The image acquisition module 8 is used to acquire a spot image containing the spot. The detection module 9 is connected to the image acquisition module and is used to detect the spot image to determine whether the through-silicon via (TSV) of the TSV wafer 200 has defects. In this embodiment, the detection module 9 is connected to the image acquisition module 8, a first driver, a second driver, and a third driver 61, and controls the operation of the first driver, the second driver, and the third driver 61. The image acquisition module 8 first acquires the spot image. The detection module 9 extracts the spot features of the spot image and generates a sharpness index. Then, based on the sharpness index, it determines whether it is necessary to adjust the parameters of the light source 1 and / or the wafer support portion 2 used to support the TSV wafer 200. If the sharpness index is lower than the target index threshold, the detection module 9 controls the operation of the first driver, the second driver, and the third driver 61 to make the sharpness index equal to or higher than the target index threshold. After the sharpness reaches the target, it determines whether the TSV of the TSV wafer 200 has defects. Here, we will not go into detail about the methods for detecting defects in through-silicon vias.

[0049] Therefore, those skilled in the art should recognize that although numerous exemplary embodiments of the present invention have been shown and described in detail herein, many other variations or modifications conforming to the principles of the present invention can be directly determined or derived from the disclosure of the present invention without departing from the spirit and scope of the invention. Thus, the scope of the present invention should be understood and construed as covering all such other variations or modifications.

Claims

1. A method of through-silicon via defect detection, the method comprising: The method comprises the following steps: obtaining a light spot image formed on an optical receiving portion after a through silicon via wafer is irradiated by a light source; extracting light spot features of the light spot image to generate a definition index; when the definition index is lower than a target index threshold, adjusting parameters of the light source and / or a wafer carrying portion for carrying the through silicon via wafer according to the definition index; when the definition index is equal to or higher than the target index threshold, performing defect detection on through silicon vias of the through silicon via wafer.

2. The method of claim 1, wherein, The step of extracting light spot features of the light spot image to generate a definition index comprises the following steps: extracting an edge sharpness feature S of the light spot in the light spot image edge , a center contrast feature C contrast , a shape roundness feature R round , and an internal gray scale uniformity feature U uniform ; a definition index for characterizing light spot definition is calculated according to the following formula, C I = w1 x S edge + w2 x C contrast + w3 x R round + w4 x U uniform ​ wherein w1 is a weight coefficient of edge sharpness feature, w2 is a weight coefficient of center contrast feature, w3 is a weight coefficient of shape roundness feature, and w4 is a weight coefficient of internal gray scale uniformity feature.

3. The method for detecting through-silicon via defects according to claim 2, characterized in that, The weight coefficient dynamic adjustment method of each light spot feature comprises the following steps: obtaining initial weight coefficients of each light spot feature; determining contribution degrees of each light spot feature to defect determination results according to real-time analysis of correlation between a plurality of light spot features and defect determination results; correcting the initial weight coefficients of the corresponding light spot features according to the contribution degrees to obtain updated weight coefficients.

4. The TSV defect detection method according to any one of claims 1-3, wherein, In the step of adjusting parameters of the light source and / or a wafer carrying portion for carrying the through silicon via wafer according to the definition index when the definition index is lower than a target index threshold, the parameters of the light source include light source brightness, position and irradiation angle, and the parameters of the wafer carrying portion include the height of the wafer carrying portion.

5. The method for detecting through-silicon via defects according to claim 4, characterized in that, The parameters of the light source and the wafer carrying portion are adjusted in the following manner: when the brightness of the light spot in the light spot image is insufficient or oversaturated, adjusting the brightness of the light source; when the edge of the light spot is blurred, adjusting the distance between the light source and the through silicon via wafer; when the shape of the light spot is offset or distorted, adjusting the irradiation angle of the light source; when the light spot image is out of focus, adjusting the height of the wafer carrying portion.

6. The method of claim 5, wherein the step of detecting the TSV defects is performed by a method comprising: The step of performing defect detection on through silicon vias of the through silicon via wafer when the definition index is equal to or higher than the target index threshold comprises the following steps: comparing the number of detected light spots with the actual number of through silicon vias to determine whether there is an unpenetrated defect in the through silicon vias; measuring the diameter and / or brightness of the light spot in the light spot image and comparing it with a preset threshold range to determine whether there is a blocked defect or a partially unpenetrated defect in the through silicon vias.

7. A through silicon via defect detection system, comprising: The method comprises the following steps: obtaining a light spot image formed on an optical receiving portion after a through silicon via wafer is irradiated by a light source; extracting light spot features of the light spot image to generate a definition index; when the definition index is lower than a target index threshold, adjusting parameters of the light source and / or a wafer carrying portion for carrying the through silicon via wafer according to the definition index; 8. The through silicon via defect detection system of claim 7, wherein, when the definition index is equal to or higher than the target index threshold, performing defect detection on through silicon vias of the through silicon via wafer. The light source is arranged to be operable or controllable to move along a vertical direction perpendicular to the wafer carrying portion; the wafer carrying portion is arranged to be operable or controllable to move along the vertical direction.

9. The through silicon via defect detection system of claim 8, wherein, The wafer carrying part has a clamping assembly arranged to be movable in a transverse direction perpendicular to the vertical direction for clamping the through-silicon via wafer.

10. The through silicon via defect detection system of claim 7, wherein, The light spot detection system comprises: An image acquisition module is configured to acquire a light spot image containing the light spot. A detection module is connected to the image acquisition module and configured to detect the light spot image to determine whether the through-silicon via wafer has a defect.

Images