Wafer defect detection method and wafer defect scanning device

Abstract

The invention provides a wafer defect detection method and a wafer defect scanning device.The wafer defect detection method is applied to the wafer defect scanning device, the wafer defect scanning device comprises a plurality of lenses, and the wafer defect detection method comprises the steps that the energy of each lens is detected, judging whether the lens is in an effective life cycle or not; scanning a wafer through the lens in the effective life cycle to obtain a scanning image of the wafer; adjusting the scanning image to obtain an adjusted image so as to highlight the difference between the defect area and the normal area; and segmenting the defect area and the normal area of the adjusted image. According to the wafer defect detection system provided by the invention, the use condition of the lens is monitored, meanwhile, the difference of different areas of images scanned by a plurality of lenses is eliminated, and the wafer defect detection efficiency and accuracy are improved.

Description

Technical Field

[0001] The present application relates to the field of semiconductor manufacturing technology, and in particular to a wafer defect detection method and a wafer defect scanning device. Background Art

[0002] In the semiconductor manufacturing process, two or more wafers often need to be bonded, and the bonded wafers need to be inspected for defects before subsequent processes are performed. Current inspection machines are generally equipped with multiple lenses to scan different inspection areas of the wafer, and finally determine the specific distribution of wafer defects based on the scanned images.

[0003] However, as the lens of the current inspection machine is used for a longer time, the energy of the ultrasonic wave emitted or received will gradually decay, and the degree of attenuation of different lenses is also different. As a result, the images presented after the multiple lenses of the inspection machine scan the wafer have large differences in parameters such as grayscale and contrast corresponding to different areas of different lenses, making it difficult to accurately determine the type and size of wafer defects. At the same time, the replacement frequency of the lens of the current inspection machine is controlled by the usage time of the lens. For lenses whose actual usable time is greater than the controlled time, the lens will not be fully used, resulting in a waste of the lens; for lenses whose actual usable time is less than the controlled time, the abnormal monitoring of the lens will not be timely, which will affect the accuracy of wafer defect detection. Summary of the invention

[0004] The present application provides a wafer defect detection method, which realizes the monitoring of the usage of the lens, while eliminating the differences between the scanned images of different lenses, thereby improving the efficiency and accuracy of wafer defect detection.

[0005] To solve the above technical problems, the present application provides a wafer defect detection method, which is applied to a wafer defect scanning device, wherein the wafer defect scanning device includes multiple lenses, and the wafer defect detection method includes: detecting the energy of each of the lenses to determine whether the lens is within its effective life cycle; scanning the wafer with the lens within its effective life cycle to obtain a scanned image of the wafer; adjusting the scanned image to obtain an adjusted image to highlight the difference between the defective area and the normal area; and segmenting the defective area and the normal area of ​​the adjusted image.

[0006] In some embodiments, detecting the energy of each lens and determining whether the lens is within a valid life cycle includes: detecting and calculating the ultrasonic energy value received by each lens after scanning the wafer; and determining whether the ultrasonic energy value received by the lens after scanning the wafer belongs to an energy value in a lens energy intensity database within a valid life cycle.

[0007] In some embodiments, the ultrasonic energy value received by the lens after scanning the wafer is calculated by the grayscale of the corresponding area to be detected when the lens scans the wafer, the number of times the lens is used, the usage time of the lens, the rated ultrasonic power of the lens, and the duration of the rated ultrasonic power of the lens.

[0008] In some embodiments, the calculation formula of the ultrasonic energy value received by the lens after scanning the wafer is:

[0009] E=(β*b+γ*T*n^3)*P*t

[0010] Among them, E is the ultrasonic energy value received after the lens scans the wafer, b is the grayscale of the corresponding area to be detected when the lens scans the wafer, n is the number of times the lens is used, T is the usage time of the lens, P is the rated ultrasonic power of the lens, t is the rated ultrasonic power action time of the lens, and β and γ are the empirical values ​​of the calculation formula.

[0011] In some embodiments, the scanning image is adjusted to obtain an adjusted image to highlight the difference between the defective area and the normal area, including: dividing the scanning image into multiple sub-scanning images, the multiple sub-scanning images respectively correspond to the scanning areas of the multiple lenses, and detecting the mean grayscale of the multiple sub-scanning images; performing image enhancement on the multiple sub-scanning images through a logarithmic transformation algorithm, adjusting the grayscale of the multiple sub-scanning images to the optimal grayscale value corresponding to the wafer, and remerging the multiple sub-scanning images; and correcting the parameters of the multiple sub-scanning images to the same value through an adaptive histogram equalization algorithm.

[0012] In some embodiments, segmenting the defective area and the normal area of ​​the adjusted image includes: obtaining a grayscale histogram of the middle area of ​​the adjusted image, and calculating a segmentation threshold of the adjusted image by the relative position of the distribution of grayscale value frequencies in the grayscale histogram to obtain the defect distribution in the middle area of ​​the wafer.

[0013] In some embodiments, the grayscale histogram of the middle area of ​​the adjustment image is obtained, and the segmentation threshold of the adjustment image is calculated by the relative position of the distribution of the grayscale value frequency in the grayscale histogram to obtain the defect distribution in the middle area of ​​the wafer, including: calculating the grayscale threshold of the defective area by the peak values ​​of multiple peaks in the grayscale histogram, the positions of the peaks, the kurtosis and the skewness, and performing threshold segmentation on the adjustment image according to the grayscale threshold.

[0014] In some embodiments, the grayscale threshold includes a first grayscale threshold and a second grayscale threshold, the grayscale value of some normal areas in the adjusted image is between the first grayscale threshold and the second grayscale threshold, the grayscale value of some first defective areas in the adjusted image is less than the first grayscale threshold, and the grayscale value of some second defective areas in the adjusted image is greater than the second grayscale threshold.

[0015] In some embodiments, segmenting the defective area and the normal area of ​​the adjusted image further includes: performing edge detection on the edge area of ​​the adjusted image to obtain defect distribution of the wafer edge area.

[0016] In some embodiments, before detecting the energy of each lens and determining whether the lens is usable, the process further includes: dividing the wafer's inspection area into a plurality of wafer sub-inspection areas, wherein the plurality of wafer sub-inspection areas respectively correspond to a plurality of scanning areas of the lenses.

[0017] In some embodiments, after segmenting the defective area and the normal area of ​​the adjusted image, the method further includes: performing image denoising on the adjusted image to remove image noise; and identifying and marking the defective area of ​​the adjusted image.

[0018] In order to solve the above technical problems, the present application also provides a wafer defect scanning device, comprising: multiple lenses for scanning a wafer to obtain a scanned image of the wafer; a processing unit connected to the lenses, and the processing unit is configured to drive the lenses to execute any of the methods described above.

[0019] The present application provides a wafer defect detection method, which is applied to a wafer defect scanning device, wherein the wafer defect scanning device includes multiple lenses, and the wafer defect detection method includes: detecting the energy of each lens to determine whether the lens is within the effective life cycle; scanning the wafer through the lens within the effective life cycle to obtain a scanned image of the wafer; adjusting the scanned image to obtain an adjusted image to highlight the difference between the defective area and the normal area; and segmenting the defective area and the normal area of ​​the adjusted image. The wafer defect detection system provided by the present application realizes the monitoring of the usage of the lens, while eliminating the difference between different areas of the image after scanning by multiple lenses, thereby improving the efficiency and accuracy of wafer defect detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0021] Figure 1 A schematic diagram of a process flow of a wafer defect detection method provided in some embodiments of the present application;

[0022] Figure 2 A schematic diagram of specific steps of a wafer defect detection method provided in some embodiments of the present application;

[0023] Figure 3a A schematic diagram of a scanned image of a wafer area to be inspected formed after scanning the wafer area to be inspected by four lenses provided in some embodiments of the present application;

[0024] Figure 3b A schematic diagram of dividing a scanned image into a plurality of sub-scanned images provided in some embodiments of the present application;

[0025] Figure 3c A schematic diagram of an enhanced image obtained by adjusting sub-scan images and merging them provided in some embodiments of the present application;

[0026] Figure 3d A schematic diagram of an adjusted image obtained after adjusting an enhanced image provided in some embodiments of the present application;

[0027] Figure 4 A schematic diagram of adjusting the grayscale histogram of an image provided in some embodiments of the present application;

[0028] Figure 5 A block diagram of a partial structure of a wafer defect scanning device provided in some embodiments of the present application. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0030] The terms "first" and "second" in this application are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0031] All directional indications in the embodiments of the present application (such as up, down, left, right, front, back, top, bottom...) are only used to explain the relative position relationship, movement status, etc. between the components in a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0032] In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0033] Unless otherwise defined, the term “substantially” mentioned in this application, with respect to numerical quantities or quantitative relationships, can be understood as a range of approximately ±15% of a certain numerical value.

[0034] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0035] Currently, defect detection for wafers cannot eliminate the differences in parameters of different areas of images scanned by different lenses, and cannot accurately monitor the usage of lenses.

[0036] In order to solve the above technical problems, the present application provides a wafer defect detection method, see Figure 1 and Figure 2 , Figure 1 A schematic diagram of a process flow of a wafer defect detection method provided in some embodiments of the present application; Figure 2 A schematic diagram of specific steps of a wafer defect detection method provided in some embodiments of the present application. The wafer defect detection method comprises:

[0037] Step S1: Detect the energy of each lens to determine whether the lens is within its effective life cycle.

[0038] Before performing wafer defect detection, the energy of multiple lenses scanning the wafer is detected and identified to determine whether the lens is within its effective life cycle, that is, to ensure that the lens has the ability to detect wafer defects. This is to avoid excessive attenuation of the ultrasonic energy emitted or received by the lens as the lens increases in use time, which causes the grayscale, signal-to-noise ratio, contrast and other parameters of the image obtained after the lens scan to decrease, ultimately leading to missed detection or false detection of wafer defects. In some embodiments, step S1 further includes:

[0039] Step S11: Detect and calculate the ultrasonic energy value received by each lens after scanning the wafer.

[0040] In some embodiments, the energy of the detected lens is the energy value of the ultrasonic wave received after the lens scans the corresponding area to be detected of the wafer. In some embodiments, the energy value of the ultrasonic wave received after the lens scans the corresponding area to be detected of the wafer is calculated by the grayscale of the lens scanning the corresponding area to be detected of the wafer, the number of times the lens is used, the usage time of the lens, the rated ultrasonic power of the lens, and the duration of the rated ultrasonic power of the lens. Specifically, the calculation formula for the energy value of the ultrasonic wave received after the lens scans the corresponding area to be detected of the wafer is constructed by algorithm fitting as follows:

[0041] E=(β*b+γ*T*n^3)*P*t

[0042] Wherein, E is the energy value of the ultrasonic wave received after the lens scans the corresponding area to be detected of the wafer, b is the grayscale of the corresponding area to be detected of the wafer scanned by the lens, n is the number of times the lens is used, T is the usage time of the lens, P is the rated ultrasonic power of the lens, t is the duration of the rated ultrasonic power of the lens, and β and γ are the empirical values ​​of the algorithm fitting model. Specifically, in some embodiments, the value range of b is grayscale level 0-255, n is an integer greater than 0, the value of T is positively correlated with n, the value of P is 50W, the value of t is 0.015s, the value of β is 0.007, and the value of γ is -1*10e-14. In some embodiments, the detected lens energy value is the average value of the ultrasonic energy of one million points received by the lens scanning the corresponding area to be detected of the wafer.

[0043] Step S12: determining whether the ultrasonic energy value received by the lens after scanning the wafer belongs to an energy value in a lens energy intensity database within a valid life cycle.

[0044] In some embodiments, the lens energy intensity database within the effective life cycle is a collection of standard values ​​of ultrasonic energy received when the scanning lens is within the effective life cycle, that is, when the scanning lens is in a normally usable state, scanning wafers of different products. That is, due to different process flows, the energy values ​​of ultrasonic waves received by the lens when scanning wafers of different products are also different. When executing step S12, it is necessary to compare the ultrasonic energy value received by the lens after scanning the wafer with the energy standard value of the wafer of the corresponding product in the lens energy intensity database within the effective life cycle, based on the wafer scanned by the lens.

[0045] When the energy value of the ultrasonic wave received after the lens scans the corresponding area to be detected of the wafer does not belong to the lens energy intensity database within the valid life cycle, it means that the lens is not within the valid life cycle, that is, the lens is in an unusable state, and the alarm system can be triggered to remind the staff to replace the lens. Further, the energy value of the ultrasonic wave received after the replaced lens scans the corresponding area to be detected of the wafer, and is compared with the energy value of the wafer of the corresponding product in the usable lens energy intensity database to determine whether the replaced lens is within the valid life cycle. When the energy value of the ultrasonic wave received after the lens scans the corresponding area to be detected of the wafer belongs to the usable lens energy intensity database, it means that the lens is in a usable state and can be used for subsequent wafer defect detection.

[0046] The wafer defect detection system provided in the present application monitors the usage of the scanning lens by detecting the energy of the scanning lens before wafer defect detection and comparing it with the energy value in the available lens energy intensity database, thereby improving the accuracy of monitoring the usability of multiple lenses, avoiding the waste of under-used lenses and the deviation of wafer defect detection due to the use of abnormal lenses, and improving the efficiency of the wafer process.

[0047] See also Figure 1 In some embodiments, the wafer defect detection method provided by the present application further includes, before executing step S1:

[0048] Step S0: dividing the wafer area to be inspected into a plurality of wafer sub-areas to be inspected, wherein the plurality of wafer sub-areas to be inspected respectively correspond to a plurality of scanning areas of the lenses.

[0049] In the process of wafer defect detection, it is often necessary to use multiple lenses to scan the wafer, so it is necessary to divide the wafer's area to be detected into multiple wafer sub-areas to be detected according to the areas scanned by the multiple lenses, so that the energy value of the ultrasonic wave received after the lens scans the wafer is subsequently detected and calculated without interference from other areas. For example, in some embodiments, wafer defect detection requires the use of 4 lenses to scan different areas of the wafer's area to be detected, then the wafer's area to be detected needs to be divided into 4 wafer sub-areas to be detected according to the scanning areas of the 4 lenses. In other embodiments, wafer defect detection can also use 3, 5 or other numbers of lenses to scan the wafer's area to be detected. The following will take the example of scanning the wafer's area to be detected using 4 lenses.

[0050] Step S2: Scanning the wafer through the usable lens to obtain a scanned image of the wafer.

[0051] See also Figure 3a , Figure 3a A schematic diagram of a scanned image of a wafer area to be inspected formed after scanning the wafer area to be inspected by four lenses provided in some embodiments of the present application. After determining that multiple lenses are usable, the wafer area to be inspected is scanned by multiple lenses, and each lens scans its corresponding wafer sub-area to be inspected, and a scanned image 100 of the wafer area to be inspected can be obtained.

[0052] Step S3: adjusting the scanned image to obtain an adjusted image to highlight the difference between the defective area and the normal area.

[0053] like Figure 3a As shown, due to the different energies of ultrasonic waves emitted and received by multiple lenses, the parameters of different areas of the image obtained after the multiple lenses scan the corresponding wafer sub-area to be inspected are quite different. Therefore, it is necessary to adjust the obtained scanned image 100 so that the grayscale of the image obtained after the multiple lenses scan the corresponding wafer sub-area to be inspected is the same, and the parameters such as contrast are in the best resolution range, so as to highlight the difference between the defective area and the normal area in the scanned image 100. Figure 3b-3d , Figure 3b A schematic diagram of dividing a scanned image into a plurality of sub-scanned images provided in some embodiments of the present application; Figure 3c A schematic diagram of an enhanced image obtained by adjusting and merging sub-scan images provided in some embodiments of the present application; Figure 3d A schematic diagram of an adjusted image obtained after adjusting an enhanced image provided in some embodiments of the present application. In some embodiments, step S3 further includes:

[0054] Step S31: dividing the scanned image into a plurality of sub-scanned images, wherein the plurality of sub-scanned images correspond to the scanning areas of the plurality of lenses respectively, and detecting the average grayscale values ​​of the plurality of sub-scanned images.

[0055] See also Figure 3b According to the scanning areas of different lenses, the scanned image 100 is divided into a plurality of sub-scanned images 110, and the plurality of sub-scanned images 110 correspond to the scanning areas of the plurality of lenses respectively. In this embodiment, the scanned image 100 needs to be divided into four sub-scanned images 110 according to the scanning areas of four lenses. In other embodiments, the scanned image 100 may also be divided into a corresponding number of sub-scanned images 110 according to other numbers of lenses. At the same time, the mean value of the grayscale of each sub-scanned image 110 is detected so that the sub-scanned image 110 can be further processed and adjusted later.

[0056] Step S32: performing image enhancement on the plurality of sub-scanning images by using a logarithmic transformation algorithm, adjusting the grayscale of the plurality of sub-scanning images to the optimal grayscale value corresponding to the wafer respectively, and re-merging the plurality of sub-scanning images.

[0057] In some embodiments, the multiple sub-scan images 110 are processed by a logarithmic transformation algorithm to improve the contrast of the darker areas of the multiple sub-scan images 110 and enhance the dark details of the multiple sub-scan images 110. The specific logarithmic transformation algorithm can refer to the existing image enhancement logarithmic transformation technology, which is not described here. In other embodiments, the darker areas of the multiple sub-scan images 110 can also be enhanced by other image processing methods. After the darker areas of the multiple sub-scan images 110 are enhanced, the grayscale of the multiple sub-scan images 110 is adjusted to the optimal grayscale value corresponding to the scanned wafer. Due to different process flows, the optimal grayscale values ​​corresponding to wafers of different products are also different. See Figure 3c After adjusting the grayscale of the plurality of sub-scan images 110 to the optimal grayscale value of the corresponding wafer, the plurality of sub-scan images 110 are merged to obtain the enhanced image 200. In some embodiments, the optimal grayscale value of the plurality of sub-scan images 110 is adjusted to 107.

[0058] Step S33: Correcting the parameters of the plurality of sub-scanned images to the same value by using an adaptive histogram equalization algorithm.

[0059] See also Figure 3c-3dIn some embodiments, the enhanced image 200 is processed by an adaptive histogram equalization algorithm (AHE) to improve the contrast of the enhanced image 200. The contrast of each area of ​​the adjusted image 300 obtained after adjustment is the same, thereby highlighting the difference between the defective area and the normal area in the adjusted image 300. The specific adaptive histogram equalization algorithm can refer to the prior art of adaptive histogram equalization, which is not described here. In other embodiments, other image processing algorithms can also be used to adjust the contrast of the enhanced image 200, which is not limited here.

[0060] Through the above-mentioned multi-step image processing algorithm, the multiple sub-scanned images 110 corresponding to the multiple lenses are first adjusted and processed separately, and then the images obtained after the multiple lenses scan the wafer's inspection area are adjusted for overall contrast balance, so that the grayscale, contrast and other parameters of the final adjusted image 300 reach the optimal inspection level corresponding to the wafer. This not only eliminates the differences in grayscale, contrast or other representations of the image caused by a single lens scanning different areas of the wafer during movement, but also eliminates the representation differences between different areas of the image obtained after multiple lenses scan, and at the same time improves the accuracy of the image after the lens scans for wafer defects, that is, greatly reduces the omission and misdetection of wafer defects.

[0061] Step S4: Segmenting the defective area and the normal area of ​​the adjusted image.

[0062] After the representations of the various regions of the adjusted image 300 are obtained to be the same through processing, the defect region and the normal region of the adjusted image 300 are segmented. Figure 1 In some embodiments, step S4 includes:

[0063] Step S41: obtaining a grayscale histogram of the middle area of ​​the adjusted image, and calculating a segmentation threshold of the adjusted image by the relative position of the distribution of grayscale value frequencies in the grayscale histogram to obtain defect distribution in the middle area of ​​the wafer.

[0064] The energy values ​​of the ultrasonic waves received by the normal area and the defective area of ​​the wafer to be inspected area when the lens scans are different, so that the defective area in the adjusted image 300 has a certain grayscale change relative to the normal area in the adjusted image 300; further, the types and sizes of the defective areas in the wafer to be inspected area are different, and the ultrasonic energy values ​​received after the lens scanning are also different, that is, the grayscale values ​​of defective areas of different types and sizes in the adjusted image 300 are also different. In some embodiments, for the middle area of ​​the adjusted image 300, the defective area and the normal area of ​​the adjusted image 300 can be segmented by obtaining the grayscale histogram of the adjusted image 300, and different types of defective areas can also be distinguished. The grayscale histogram of the adjusted image 300 is obtained by converting the grayscale value of each pixel in the adjusted image 300. See. Figure 4 , Figure 4 A schematic diagram of a grayscale histogram of an adjustment image provided for some embodiments of the present application. The X-axis of the grayscale histogram 400 represents the grayscale value of a pixel in the adjustment image 300, and the Y-axis represents the number of pixels in the adjustment image 300. The position of the highest peak is determined by the peak values ​​of multiple peaks in the grayscale histogram 400, as well as the kurtosis and skewness of each peak, so as to calculate the grayscale threshold of the defective area in the adjustment image 300, and the defective area and the normal area in the adjustment image 300 can be segmented by the grayscale threshold. In some embodiments, the defective area of ​​the wafer to be detected includes two types of defects, namely, bubble defects and residual defects. Among them, the bubble defect is the bubble generated between the two wafers after the wafers are bonded, and the residual defect is the residue of other substances such as organic matter on the surface of the wafer. In some embodiments, the grayscale threshold of the grayscale histogram 400 includes a first grayscale threshold Tv1 and a second grayscale threshold Tv2, which correspond to the bubble defect area and the residual defect area in the adjustment image 300, respectively. Specifically, the grayscale histogram 400 includes n peaks, and the calculation formulas of the first grayscale threshold Tv1 and the second grayscale threshold Tv2 are respectively:

[0065] Tv1=L(x+1)–a*K(x+1)+b*S(x+1)

[0066] Tv2=L(x-1)+a*K(x-1)-b*S(x-1)

[0067] Wherein, M1, ..., Mn are the peak values ​​of n peaks in the grayscale histogram 400, L1, ..., Ln are the positions of the peak values ​​of n peaks in the grayscale histogram 400, K1, ..., Kn are the kurtosis of n peaks in the grayscale histogram 400, S1, ..., Sn are the skewness of n peaks in the grayscale histogram 400, Mx is the peak value of the largest peak in the grayscale histogram 400, Lx is the position of the peak value of the largest peak in the grayscale histogram 400, Kx is the kurtosis of the largest peak in the grayscale histogram 400, and Sx is the skewness of the largest peak in the grayscale histogram 400. In some embodiments, the value range of L1, ..., Ln is grayscale level 0-255, the value of a is 3.5, and the value of b is 8.3.

[0068] The calculation formula of kurtosis K1,…,Kn is:

[0069] Kurt(X)=E[((X-μ) / σ) 4 ]=E[(X-μ) 4 ] / E[(X-μ) 2 ]) 2

[0070] The calculation formula of skewness S1,…,Sn is:

[0071] Skew(X)=E[((X-μ) / σ)]=k3 / σ3=k3 / k2 3 / 2

[0072] In some embodiments, the grayscale values ​​of some pixels in the normal area of ​​the adjusted image 300 are between the first grayscale threshold Tv1 and the second grayscale threshold Tv2, and the pixels corresponding to the grayscale values ​​between the first grayscale threshold Tv1 and the second grayscale threshold Tv2 are all pixels in the normal area; the grayscale values ​​of some pixels in the residual defect area of ​​the adjusted image 300 are less than the first grayscale threshold Tv1, and the grayscale values ​​of some pixels in the bubble defect area of ​​the adjusted image 300 are greater than the second grayscale threshold Tv2. In this way, the normal area and the defect area with similar grayscale range in the adjusted image 300 can be segmented, that is, the defect area in the adjusted image 300 that is easily confused with the normal area is segmented, and two different defect areas in the adjusted image 300 are also preliminarily distinguished. In other embodiments, when the area to be detected of the wafer includes multiple other defects, the adjusted image 300 can also be segmented by calculating the grayscale thresholds of different defect areas in the adjusted image 300 after obtaining the grayscale histogram 400.

[0073] See also Figure 1 Since the above method of obtaining the grayscale histogram for threshold segmentation has a low defect detection accuracy for the edge area of ​​the adjusted image 300, in some embodiments, step S4 further includes:

[0074] Step S42: performing edge detection on the edge area of ​​the adjusted image to obtain defect distribution of the edge area of ​​the wafer.

[0075] In some embodiments, for the edge area of ​​the adjustment image 300, the defective area in the adjustment image 300 can be detected and identified by edge detection to improve the accuracy of defect detection at the edge of the wafer area to be detected. The specific edge detection method can refer to the existing technology of image processing edge detection, which is not described here. In some embodiments, the width of the edge area of ​​the wafer for defective area detection and identification by edge detection is 1.5% of the diameter of the wafer area to be detected. In other embodiments, other areas of the wafer area to be detected can also be detected separately by edge detection and grayscale histogram threshold segmentation methods according to different situations, which are not limited here.

[0076] See also Figure 1 In some embodiments, the wafer defect detection method provided by the present application further includes, after executing step S4:

[0077] Step S5: performing image denoising on the adjusted image to remove image noise.

[0078] In some embodiments, the image 300 obtained after the above adjustment after the image of the area to be detected of the wafer scanned by multiple lenses also includes some image noise. When the adjusted image 300 is subjected to histogram threshold segmentation or edge detection defect recognition, these image noises will be segmented into the defect area of ​​the adjusted image 300, resulting in image noise in the defect area segmented by the adjusted image 300. Therefore, it is necessary to perform image denoising on the adjusted image 300 to remove image noise and improve the accuracy of wafer defect detection. In some embodiments, an image denoising algorithm is used to perform image denoising on the adjusted image 300, and the image denoising algorithm includes image algorithms such as opening operation, closing operation, contour detection, circle recognition, and right angle recognition. The specific image denoising algorithm can refer to the image opening operation, image closing operation, contour detection algorithm, circle recognition algorithm, and right angle recognition algorithm in the prior art, which will not be described here. In other embodiments, other image denoising methods can also be used to perform image denoising on the adjusted image 300, which is not limited here. After image denoising on the adjusted image 300, the segmentation of the defect area and the normal area can be finally completed.

[0079] See also Figure 1 In some embodiments, the wafer defect detection method provided by the present application further includes, after executing step S5:

[0080] Step S6: Identify and mark the defective area of ​​the adjusted image.

[0081] After the defect segmentation and denoising of the adjusted image 300, the defect area in the adjusted image 300 can be identified and marked. The defect area in the adjusted image 300 corresponds to the defect area of ​​the wafer to be inspected, thereby realizing the detection and identification of the defect area of ​​the wafer to be inspected.

[0082] In order to solve the above technical problems, the present application also provides a wafer defect scanning device for detecting and identifying wafer defects. Figure 5 , Figure 5 A block diagram of a partial structure of a wafer defect scanning device provided for some embodiments of the present application. The wafer defect scanning device 10 includes a plurality of lenses 11, which are used to scan the area to be detected of the wafer to obtain a scanning image 100 corresponding to the area to be detected of the wafer. In some embodiments, the wafer defect scanning device 10 is provided with four lenses 11 for respectively scanning different position areas of the area to be detected of the wafer; in other embodiments, the wafer defect scanning device 10 may also be provided with other numbers of lenses 11, which are not limited here. Furthermore, the wafer defect scanning device 10 also includes a processing unit 12, which is connected to the plurality of lenses 11 of the wafer defect scanning device. The processing unit 12 is used to drive the plurality of lenses 11 of the wafer defect scanning device to execute the above-mentioned wafer defect detection method.

[0083] The present application provides a wafer defect detection method, which is applied to a wafer defect scanning device 10, wherein the wafer defect scanning device 10 includes a plurality of lenses 11, and the wafer defect detection method includes: detecting the energy of each lens 11 to determine whether the lens 11 is usable; scanning the wafer through the usable lens 11 to obtain a scanned image 100 of the wafer; adjusting the scanned image 100 to obtain an adjusted image 300 to highlight the difference between the defective area and the normal area; and segmenting the defective area and the normal area of ​​the adjusted image 300. The wafer defect detection method provided by the present application realizes the monitoring of the use of the lens 11, and at the same time eliminates the difference between different areas of the image after scanning by multiple lenses 11, thereby improving the efficiency and accuracy of wafer defect detection.

[0084] The above are only implementation methods of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A wafer defect detection method, applied to a wafer defect scanning device, wherein the wafer defect scanning device comprises a plurality of lenses, characterized in that: The wafer defect detection method comprises: Detecting the energy of each lens to determine whether the lens is within its effective life cycle; Scanning a wafer through the lens within its effective life cycle to obtain a scanned image of the wafer; Adjusting the scanned image to obtain an adjusted image to highlight the difference between the defective area and the normal area; The defective area and the normal area of ​​the adjusted image are segmented.

2. The wafer defect detection method according to claim 1, characterized in that: The detecting the energy of each lens and judging whether the lens is within the valid life cycle includes: Detecting and calculating the ultrasonic energy value received by each lens after scanning the wafer; It is determined whether the ultrasonic energy value received by the lens after scanning the wafer belongs to an energy value in a lens energy intensity database within a valid life cycle.

3. The wafer defect detection method according to claim 2, characterized in that: The ultrasonic energy value received by the lens after scanning the wafer is calculated by the grayscale of the corresponding area to be detected when the lens scans the wafer, the number of times the lens is used, the usage time of the lens, the rated ultrasonic power of the lens and the duration of the rated ultrasonic power action of the lens.

4. The wafer defect detection method according to claim 3, characterized in that: The calculation formula of the ultrasonic energy value received by the lens after scanning the wafer is: E=(β*b+γ*T*n^3)*P*t Among them, E is the ultrasonic energy value received after the lens scans the wafer, b is the grayscale of the corresponding area to be detected when the lens scans the wafer, n is the number of times the lens is used, T is the usage time of the lens, P is the rated ultrasonic power of the lens, t is the rated ultrasonic power action time of the lens, and β and γ are the empirical values ​​of the calculation formula.

5. The wafer defect detection method according to claim 1, characterized in that: The step of adjusting the scanned image to obtain an adjusted image to highlight the difference between the defective area and the normal area includes: Dividing the scanned image into a plurality of sub-scanned images, wherein the plurality of sub-scanned images respectively correspond to the scanning areas of the plurality of lenses, and detecting the mean grayscale values ​​of the plurality of sub-scanned images; Performing image enhancement on the multiple sub-scanning images by using a logarithmic transformation algorithm, adjusting the grayscale of the multiple sub-scanning images to the optimal grayscale value corresponding to the wafer respectively, and re-merging the multiple sub-scanning images; The parameters of the plurality of sub-scanned images are corrected to the same value by using an adaptive histogram equalization algorithm.

6. The wafer defect detection method according to claim 1, characterized in that: The segmentation of the defective area and the normal area of ​​the adjusted image includes: obtaining a grayscale histogram of the middle area of ​​the adjusted image, and calculating a segmentation threshold of the adjusted image by the relative position of the distribution of grayscale value frequencies in the grayscale histogram to obtain the defect distribution in the middle area of ​​the wafer.

7. The wafer defect detection method according to claim 6, characterized in that: The step of obtaining a grayscale histogram of the middle area of ​​the adjusted image and calculating a segmentation threshold of the adjusted image by the relative position of the distribution of grayscale value frequencies in the grayscale histogram to obtain defect distribution in the middle area of ​​the wafer includes: The grayscale threshold of the defective area is calculated by using the peak values, positions, kurtosis and skewness of the multiple peaks in the grayscale histogram, and the adjusted image is threshold segmented according to the grayscale threshold.

8. The wafer defect detection method according to claim 7, characterized in that: The grayscale threshold includes a first grayscale threshold and a second grayscale threshold, the grayscale value of some normal areas in the adjusted image is between the first grayscale threshold and the second grayscale threshold, the grayscale value of some first defective areas in the adjusted image is less than the first grayscale threshold, and the grayscale value of some second defective areas in the adjusted image is greater than the second grayscale threshold.

9. The wafer defect detection method according to claim 6, characterized in that: The segmenting of the defective area and the normal area of ​​the adjusted image further includes: The defect distribution of the edge area of ​​the wafer is obtained by performing edge detection on the edge area of ​​the adjusted image.

10. The wafer defect detection method according to claim 1, characterized in that: Before detecting the energy of each lens and determining whether the lens is usable, the method further includes: The wafer area to be inspected is divided into a plurality of wafer sub-areas to be inspected, and the plurality of wafer sub-areas to be inspected respectively correspond to the scanning areas of the plurality of lenses.

11. The wafer defect detection method according to claim 1, characterized in that: After segmenting the defective area and the normal area of ​​the adjusted image, the method further includes: Performing image denoising on the adjusted image to remove image noise; The defective area of ​​the adjusted image is identified and marked.

12. A wafer defect scanning device, characterized in that: include: A plurality of lenses, used for scanning the wafer to obtain a scanned image of the wafer; A processing unit is connected to the lens, and the processing unit is configured to drive the lens to execute the method according to any one of claims 1-11.

Images