A method and apparatus for detecting a cop defect

Abstract

The application provides a COP defect detection method and device, and belongs to the technical field of semiconductor manufacturing. The method comprises the following steps: scanning and detecting a to-be-detected wafer to obtain a detection result of each defect point on the to-be-detected wafer; determining a first score value of a first defect point as a crystal original particle COP defect based on a multi-channel light scattering feature vector; determining a second score value of the COP defect in the to-be-detected wafer based on a proportion of the number of the first type of defect points on the to-be-detected wafer; determining a third score value of the COP defect in the to-be-detected wafer based on the spatial distribution of each defect point on the to-be-detected wafer; and determining a COP defect result of the to-be-detected wafer according to the first score value, the second score value and the third score value. The to-be-detected wafer is comprehensively scored from multiple dimensions, and finally, the COP defect detection result is determined, the accurate identification of the COP defect of a single wafer is realized, and the reliability of the detection result is improved.

Description

Technical Field

[0001] This invention relates to the field of semiconductor manufacturing technology, and in particular to a method and apparatus for detecting COP defects. Background Technology

[0002] In the semiconductor silicon wafer manufacturing process, particle counters are typically used to detect wafer surface defects based on laser scattering signals. However, when there are slightly aggregated crystal-originated particles (COP) on the wafer surface, they are often difficult to distinguish from ordinary light point defects (LPD). LPD defects can be removed through subsequent cleaning; however, COP defects are LPD-N defects, meaning they cannot be removed through cleaning. If COP defects are not detected and intercepted in time, they will affect the quality of subsequent products.

[0003] Currently, the method usually relies on the overlay analysis of the entire batch (lot) to determine the type of COP defect, but it has poor ability to identify COP defects on a single wafer. Furthermore, detection based on particle counters is prone to mislabeling slightly clustered COP defects as LPD defects, leading to missed detections.

[0004] It is evident that existing technologies suffer from poor accuracy in detecting COP defects on single wafers. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a COP defect detection method and apparatus to solve the problem of poor accuracy in COP defect detection for single wafers in the prior art.

[0006] To achieve the above objectives, the technical solution adopted in the embodiments of the present invention is as follows:

[0007] This invention provides a COP defect detection method, comprising:

[0008] The wafer under test is scanned and inspected to obtain the detection results of each defect point on the wafer under test. The detection results include the multi-channel light scattering feature vector, position information and optically determined defect type of each defect point. The multi-channel light scattering feature vector is used to represent the optical scattering characteristics of the defect point in different detection channels. The optically determined defect type includes adhesive defect type and non-adhesive defect type.

[0009] Based on the multi-channel light scattering feature vector, a first score value is determined for the first defect point to be a crystal native particle COP defect. The first defect point is any defect point on the wafer under test. The first score value is used to quantify the probability that the first defect point is a COP defect based on the dimension of the multi-channel light scattering feature of the defect point.

[0010] Based on the proportion of the first type of defect points on the wafer under test, a second score value is determined for the presence of COP defects in the wafer under test. The first type is the non-adhesive defect type indicated in the optically determined defect type. The second score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the proportion of defect points corresponding to the defect type.

[0011] Based on the spatial distribution of each defect point on the wafer under test, a third score value is determined for the presence of COP defects in the wafer under test. The spatial distribution is determined according to the position information of each defect point in the detection results. The third score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the dimension of the spatial distribution of the defect points.

[0012] The COP defect result of the wafer under test is determined based on the first score, the second score, and the third score.

[0013] In some embodiments, determining the first score value of the first defect point as a crystal native particle COP defect based on the multi-channel light scattering feature vector includes:

[0014] Obtain the multi-channel light scattering feature vector of the sample defect point, wherein the sample defect point is a predetermined non-adhesive defect point;

[0015] Calculate the first similarity between the multi-channel light scattering feature vector of the first defect point and the multi-channel light scattering feature vector of the sample defect point;

[0016] The first similarity is converted into a score value to obtain a first score value, which is proportional to the first similarity.

[0017] In some embodiments, determining a second score value for the presence of COP defects in the wafer under test based on the proportion of first-type defects on the wafer under test includes:

[0018] The defect type of each defect point in the detection results is analyzed to obtain the number of defect points of the first type and the number of defect points of the second type, and the adhesion defect type indicated in the optically determined defect type of the second type.

[0019] The proportion of the first type of defect points is determined based on the number of the first type of defect points and the number of the second type of defect points;

[0020] The percentage of defects of the first type is converted into a score value to obtain the second score value, which is proportional to the percentage of defects of the first type.

[0021] In some embodiments, determining the third score value for the presence of COP defects in the wafer under test based on the spatial distribution of each defect point on the wafer under test includes:

[0022] Based on the location information of each defect point in the detection results, the target spatial distribution characteristics of each defect point on the wafer under test are determined;

[0023] Calculate the second similarity between the target spatial distribution feature and the first spatial distribution feature, the third similarity between the target spatial distribution feature and the second spatial distribution feature, and the fourth similarity between the target spatial distribution feature and the third spatial distribution feature. The first spatial distribution feature indicates that the defect points only gather in the central region of the wafer, the second spatial distribution feature indicates that the defect points only gather in the edge region of the wafer, and the third spatial distribution feature indicates that the defect points gather in both the central region and the edge region of the wafer.

[0024] The fifth similarity is converted into a score value to obtain a third score value, which is proportional to the fifth similarity. The fifth similarity is the maximum value among the second, third, and fourth similarities.

[0025] In some embodiments, determining the COP defect result of the wafer under test based on the first score value, the second score value, and the third score value includes:

[0026] A comprehensive score is calculated based on the first score, the second score, the third score, the first weight parameter corresponding to the first score, the second weight parameter corresponding to the second score, and the third weight parameter corresponding to the third score.

[0027] If the overall score is greater than or equal to a preset score threshold, it is determined that a COP defect exists in the wafer under test.

[0028] In some embodiments, the step of scanning and inspecting the wafer under test to obtain the detection result of each defect point on the wafer under test includes:

[0029] The wafer under test is scanned and inspected using a particle counter to obtain the Klarf file output by the particle counter and the multi-channel light scattering signal of each defect point on the wafer under test in different detection channels. The detection channels include a normal incidence channel, an oblique incidence channel and a bright field detection channel.

[0030] The Klarf file is parsed to obtain the location information of each defect point on the wafer under test and the defect type determined by optical analysis;

[0031] Feature extraction is performed on the multi-channel light scattering signal to obtain the multi-channel light scattering feature vector.

[0032] This invention also provides a COP defect detection device, comprising:

[0033] The detection module is used to scan and detect the wafer under test, and obtain the detection results of each defect point on the wafer under test. The detection results include the multi-channel light scattering feature vector, position information and optically determined defect type of each defect point. The multi-channel light scattering feature vector is used to represent the optical scattering characteristics of the defect point in different detection channels. The optically determined defect type includes adhesive defect type and non-adhesive defect type.

[0034] The first determining module is used to determine a first score value for a first defect point as a crystal native particle COP defect based on the multi-channel light scattering feature vector. The first defect point is any defect point on the wafer under test. The first score value is used to quantify the probability that the first defect point is a COP defect based on the dimension of the multi-channel light scattering feature of the defect point.

[0035] The second determining module is used to determine a second score value for the presence of COP defects in the wafer under test based on the proportion of the number of first type defect points on the wafer under test. The first type is the non-adhesive defect type indicated in the optically determined defect type. The second score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the dimension of the proportion of defect points corresponding to the defect type.

[0036] The third determining module is used to determine a third score value for the presence of COP defects in the wafer under test based on the spatial distribution of each defect point on the wafer under test. The spatial distribution is determined according to the position information of each defect point in the detection result. The third score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the dimension of the spatial distribution of the defect points.

[0037] The fourth determining module is used to determine the COP defect result of the wafer under test based on the first score value, the second score value and the third score value.

[0038] This invention also provides an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the COP defect detection method as described above.

[0039] This invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the COP defect detection method described above.

[0040] This invention also provides a computer program product, including computer instructions that, when executed by a processor, implement the steps of the COP defect detection method as described above.

[0041] The beneficial effects of this invention are:

[0042] In this embodiment, the wafer under test is scanned to obtain the multi-channel light scattering feature vector of each defect point on the wafer under test and the basic detection results. Then, based on the multi-channel light scattering feature vector, the probability of COP of a single defect point is initially judged at the optical response level, but it is not directly used as the COP detection result of the wafer under test. Furthermore, the wafer under test is comprehensively scored from the dimensions of the multi-channel light scattering feature vector, the dimensions of the global defect type composition, and the dimensions of the spatial distribution of each defect point, in combination with the proportion of the first type of defect points on the wafer under test and the spatial distribution of each defect point. Finally, the COP defect detection result is determined, so as to achieve accurate identification of COP defects of a single wafer and improve the reliability of the detection results. Attached Figure Description

[0043] Figure 1 This is a flowchart illustrating the COP defect detection method in an embodiment of the present invention.

[0044] Figure 2 This image shows the microscopic morphology of COP defects in an embodiment of the present invention.

[0045] Figure 3 This is a schematic diagram of the COP defect detection device in an embodiment of the present invention;

[0046] Figure 4 This is a schematic diagram of the structure of the electronic device in an embodiment of the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0048] In the semiconductor silicon wafer manufacturing process, a particle counter can scan and inspect the wafer using laser scattering signals and generate a Klarf file. However, in actual inspection of a single wafer, when there are some slightly clustered COP defects on the wafer surface, these slightly clustered COP defects are easily mistakenly marked as LPD defects. This invention provides a COP defect detection method and apparatus that can improve the accuracy of COP defect detection for single wafers.

[0049] This invention provides a method for detecting COP defects, such as... Figure 1 As shown, it includes the following steps:

[0050] Step 101: Scan and inspect the wafer under test to obtain the detection results of each defect point on the wafer under test. The detection results include the multi-channel light scattering feature vector, position information and optically determined defect type of each defect point. The multi-channel light scattering feature vector is used to represent the optical scattering characteristics of the defect point in different detection channels. The optically determined defect type includes adhesive defect type and non-adhesive defect type.

[0051] In this step, the surface of the wafer under test is scanned by laser to obtain the multi-channel light scattering feature vector, location information, and optically determined defect type of each defect point on the wafer, providing data support for further accurate analysis.

[0052] Among them, the multi-channel light scattering feature vector can reflect the comprehensive optical scattering behavior of the defect point in different detection channels. Therefore, by comparing the similarity between the current defect point and the sample defect point with specific defect characteristics in the dimension of the multi-channel light scattering feature vector, the defect type of the current defect point can be preliminarily determined, thereby preliminarily determining the COP detection result of the wafer under test.

[0053] The location information and optically determined defect types can be basic detection results directly output by relevant wafer defect detection equipment or systems. These results include the location information of each defect point and the optically determined defect type. Based on the location information of the defect points, the spatial distribution of the defect points can be determined. By determining the defect type from the dimension of the spatial distribution of the defect points, the COP detection result of the wafer under test can be further determined. The optically determined defect types can include LPD defects (i.e., adhesion defects) and LPD-N defects (i.e., non-adhesion defects) directly marked by the relevant wafer defect detection equipment or system. LPD defects can be removed by subsequent cleaning, while LPD-N defects cannot. Considering that optically determined defect types can easily mislabel slightly clustered COP defects as LPD defects, in this embodiment, the ratio of LPD defects to LPD-N defects can be determined based on the optically determined defect types. The COP detection result of the wafer under test is determined from the dimension of the global defect type composition. The microstructure of the COP defects, such as... Figure 2 As shown.

[0054] The detection results for each defect point on the wafer under test are obtained, as detailed in the following description:

[0055] In some embodiments, the step of scanning and inspecting the wafer under test to obtain the detection result of each defect point on the wafer under test includes:

[0056] The wafer under test is scanned and detected using a particle counter to obtain the Klarf file output by the particle counter and the multi-channel light scattering signal of each defect point on the wafer under test in different detection channels. The detection channels include a normal incidence channel, an oblique incidence channel and a bright field detection channel.

[0057] The Klarf file is parsed to obtain the location information of each defect point on the wafer under test and the defect type determined by optical analysis;

[0058] Feature extraction is performed on the multi-channel light scattering signal to obtain the multi-channel light scattering feature vector.

[0059] In this embodiment, a Particle Counter can be used to scan and inspect the wafer under test. The defect result file (i.e., a Klarf file) output by the Particle Counter and the multi-channel light scattering signals obtained during the scanning of the wafer surface using the laser scattering principle are acquired. Specifically, for each defect point, light scattering signals from multiple detection channels are simultaneously acquired during the same scan. These detection channels may include: a normal illumination channel, an oblique illumination channel, and a bright field detection channel, etc. This acquisition method, using laser scattering from different angles and optical paths, can comprehensively capture the optical characteristics of defects, reducing the likelihood of missed detections or false positives that can occur with a single channel.

[0060] Then, feature extraction is performed on the acquired multi-channel light scattering signals. For example, the scattering intensity values ​​of the defect point in each channel and the ratio of scattering intensity between different channels can be extracted and then combined in a preset order to construct the multi-channel light scattering feature vector corresponding to the defect point. The multi-channel light scattering feature vector reflects the comprehensive optical scattering behavior of the defect point in different detection channels, and the defect type of the current defect point is initially determined, thereby initially determining the COP detection result of the wafer under test. See step 102 for details.

[0061] Furthermore, the Klarf file is parsed to obtain the location information of each defect point and the defect type determined by optical analysis. Thus, the spatial distribution of defect points can be determined based on their location information. By determining the defect type from the dimension of the spatial distribution of defect points, the COP detection result of the wafer under test can be further determined, as detailed in subsequent step 103. The ratio of LPD defects to LPD-N defects can be determined based on the defect type determined by optical analysis. The COP detection result of the wafer under test can be determined from the dimension of the global defect type composition, as detailed in subsequent step 104.

[0062] Step 102: Based on the multi-channel light scattering feature vector, determine the first score value of the first defect point as a crystal native particle COP defect. The first defect point is any defect point on the wafer to be tested. The first score value is used to quantify the probability that the first defect point is a COP defect based on the dimension of the multi-channel light scattering feature of the defect point.

[0063] In this step, COP, as a native pit defect in the crystal, has specific multi-channel light scattering feature vectors from multiple detection channels, such as the scattering signal characteristics of the normally incident channel and a fixed range of intensity ratios between different channels. The similarity between the multi-channel light scattering feature vector corresponding to any defect point on the wafer under test and the multi-channel light scattering feature vector corresponding to a sample defect point can be calculated to determine the first score value for the current defect point as a COP defect. The higher the first score value, the greater the probability that the current defect point is an LPD-N defect, i.e., the greater the probability that the current defect point is a COP defect. Thus, the first score value can preliminarily determine whether the current defect is closer to an attached defect type or a non-attached defect type at the optical response level. However, this step is not directly used as the basis for determining the COP detection result; it needs to be further combined with subsequent steps for comprehensive judgment to improve the accuracy of COP defect detection results for single wafers. The process of obtaining the first score value can be specifically described below:

[0064] In some embodiments, determining the first score value of the first defect point as a crystal native particle COP defect based on the multi-channel light scattering feature vector includes:

[0065] Obtain the multi-channel light scattering feature vector of the sample defect point, wherein the sample defect point is a predetermined non-adhesive defect point;

[0066] Calculate the first similarity between the multi-channel light scattering feature vector of the first defect point and the multi-channel light scattering feature vector of the sample defect point;

[0067] The first similarity is converted into a score value to obtain a first score value, which is proportional to the first similarity.

[0068] In this embodiment, the multi-channel light scattering feature vectors corresponding to different sample defect points are first calibrated manually, by equipment, or by a system to construct a standard optical feature library. This standard optical feature library includes multi-channel light scattering feature vectors corresponding to non-adhesive defect types (i.e., LPD-N defects). After obtaining the multi-channel light scattering feature vectors of the sample defect points, a vector similarity algorithm can be used to calculate the first similarity between the multi-channel light scattering feature vector of the first defect point and the multi-channel light scattering feature vector of the sample defect point. Then, the first similarity is mapped to a score value to obtain the first score value. Thus, the higher the first similarity, the higher the matching degree between the optical scattering behavior of the first defect point and the optical scattering behavior of non-adhesive defect types in the standard optical feature library. Therefore, the probability that the first defect point is a COP is greater, and the first score value is also higher. By comparing the similarity between the current defect point and sample defect points with non-adhesive defect types in the dimension of multi-channel light scattering feature vectors and mapping it to a score, the first score is used to preliminarily determine whether the current defect is closer to an adhesive defect type or a non-adhesive defect type in terms of optical response, quantifying the probability that the current defect is a COP. However, this score is not directly used as the basis for determining the COP detection result. It can be further combined with subsequent steps for comprehensive judgment to improve the accuracy of COP defect detection results for single wafers.

[0069] Step 103: Based on the proportion of the number of first type defects on the wafer under test, determine a second score value for the presence of COP defects in the wafer under test. The first type is the non-adhesive defect type indicated in the optically determined defect type. The second score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the proportion of defect types corresponding to defect points.

[0070] In this step, after initially determining the defect type of the defect points on the wafer under test through the above steps, and thus initially determining the COP detection result of the wafer under test, the defect type of each defect point can be analyzed to determine the proportion of the first type (i.e., LPD-N defect) defect points on the wafer under test. This determines the second score value for the presence of COP defects in the wafer under test. From the dimension of the proportion of defect types corresponding to defect points, i.e., the dimension of the global defect type composition, the probability of the presence of COP defects in the wafer under test is quantified. Combining the aforementioned first score value, the second score value obtained in this step, and the third score value obtained in subsequent steps, the COP defect detection result can be comprehensively determined from multi-channel light scattering characteristics, the proportion of defect types corresponding to defect points, and the spatial distribution of defect points. This achieves accurate identification of COP defects on a single wafer, reduces the over-detection rate, and improves the reliability of the detection results. The process of obtaining the second score value can be specifically described as follows:

[0071] In some embodiments, determining a second score value for the presence of COP defects in the wafer under test based on the proportion of first-type defects on the wafer under test includes:

[0072] The defect type of each defect point in the detection results is analyzed to obtain the number of defect points of the first type and the number of defect points of the second type, and the adhesion defect type indicated in the optically determined defect type of the second type.

[0073] The proportion of the first type of defect points is determined based on the number of the first type of defect points and the number of the second type of defect points;

[0074] The percentage of defects of the first type is converted into a score value to obtain the second score value, which is proportional to the percentage of defects of the first type.

[0075] In this embodiment, the defect type of each defect point in the detection results is first analyzed to count the number of defect points of the first type and the number of defect points of the second type. Among them, the first type of defect is a non-adhesive defect (i.e., LPD-N defect, COP belongs to LPD-N defect, which is a defect that cannot be cleaned and removed); the second type of defect is an adhesive defect (i.e., LPD defect, which is a defect that can be cleaned and removed).

[0076] Then, based on the number of first-type and second-type defects, the proportion of first-type defects in the wafer under test is determined, denoted as R. R can be expressed as: R = LPD - N / (LPD + LPD - N). When R is higher than a preset threshold (e.g., 0.6), it indicates an abnormally high proportion of non-adhesive defects in the wafer under test, showing a significant correlation with COP-defect wafers in historical data. The proportion of first-type defects, R, can be converted into a score value, resulting in a second score value, which is proportional to R. Thus, by statistically analyzing the ratio of LPD-N defects to LPN defects on the wafer under test and mapping it to the second score, the probability of COP defects on the entire wafer under test is quantified. This allows for a determination of whether COP is the dominant defect from the perspective of global defect type composition, further confirming the COP detection results of the wafer under test. Further comprehensive judgment is then performed in conjunction with subsequent steps to improve the accuracy of COP defect detection results for a single wafer.

[0077] Step 104: Based on the spatial distribution of each defect point on the wafer under test, determine the third score value of the presence of COP defects in the wafer under test. The spatial distribution is determined according to the position information of each defect point in the detection results. The third score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the dimension of the spatial distribution of defect points.

[0078] In this step, after initially determining the COP detection result of the wafer under test based on the dimension of the multi-channel light scattering feature vector in the previous step, and further determining the COP detection result based on the dimension of the global defect type composition, a third score value for the presence of COP defects in the wafer under test can be determined based on the spatial distribution of each defect point on the wafer under test. This further determines the COP detection result of the wafer under test from the dimension of the spatial distribution of defect points. Combining the aforementioned first score value, second score value, and the third score value obtained in this step, the COP defect detection result can be comprehensively determined from the multi-channel light scattering features, the proportion of defect types corresponding to defect points (i.e., LPD-N defect proportion), and the spatial distribution of defect points. This achieves accurate identification of COP defects on a single wafer, reduces the over-detection rate, and improves the reliability of the detection results. The process of obtaining the third score value can be specifically described as follows:

[0079] In some embodiments, determining the third score value for the presence of COP defects in the wafer under test based on the spatial distribution of each defect point on the wafer under test includes:

[0080] Based on the location information of each defect point in the detection results, the target spatial distribution characteristics of each defect point on the wafer under test are determined;

[0081] Calculate the second similarity between the target spatial distribution feature and the first spatial distribution feature, the third similarity between the target spatial distribution feature and the second spatial distribution feature, and the fourth similarity between the target spatial distribution feature and the third spatial distribution feature. The first spatial distribution feature indicates that the defect points only gather in the central region of the wafer, the second spatial distribution feature indicates that the defect points only gather in the edge region of the wafer, and the third spatial distribution feature indicates that the defect points gather in both the central region and the edge region of the wafer.

[0082] The fifth similarity is converted into a score value to obtain a third score value, which is proportional to the fifth similarity. The fifth similarity is the maximum value among the second, third, and fourth similarities.

[0083] In this embodiment, firstly, the first spatial distribution characteristics, the second spatial distribution characteristics, and the third spatial distribution characteristics can be determined based on the typical distribution pattern of COP defects. See the following description for details:

[0084] In one example, a typical distribution pattern of COP defects can be a center-clustering pattern: within a wafer center radius R1, the density of defect points is significantly higher than the average density across the entire wafer. Based on this, a first spatial distribution characteristic is determined according to the center-clustering pattern.

[0085] In one example, a typical distribution pattern of COP defects can be an edge-ring clustering pattern: a ring-shaped anomalous distribution is formed within a radius range R2 to R3 near the wafer edge. Based on this, a second spatial distribution characteristic is determined according to the edge-ring clustering pattern.

[0086] In one example, a typical distribution pattern of COP defects can be a composite clustering pattern: anomalies appear simultaneously in the central and peripheral regions. Based on this, the distribution characteristics of the third space are determined according to the composite clustering pattern.

[0087] Thus, based on the X-axis and Y-axis coordinates of each defect point parsed from the Klarf file, after determining the target spatial distribution characteristics of all defects on the wafer under test, the second similarity between the target spatial distribution characteristics and the first spatial distribution characteristics, the third similarity between the target spatial distribution characteristics and the second spatial distribution characteristics, and the fourth similarity between the target spatial distribution characteristics and the third spatial distribution characteristics are calculated respectively. The maximum similarity is taken as the fifth similarity and converted into a third score value, quantifying the probability of COP defects on the entire wafer under test. Determining the COP detection results of the wafer under test from the dimension of defect point spatial distribution further improves the accuracy of COP defect detection results for single wafers.

[0088] Step 105: Determine the COP defect result of the wafer under test based on the first score, the second score, and the third score.

[0089] In this step, a comprehensive calculation is performed based on the first score corresponding to a single-point optical feature, the second score corresponding to the global LPD-N defect ratio, and the third score corresponding to the defect spatial distribution pattern to determine the COP defect result of the wafer under test. For example, the comprehensive score obtained by comprehensively calculating the first, second, and third scores is greater than or equal to a preset score, indicating that there is a risk of COP defects in the wafer under test. This multi-dimensional joint analysis can effectively distinguish COP from ordinary LPD, improving the accuracy of the detection results and reducing misjudgments and scrap. Specifically, a multi-weighted comprehensive score can be used to determine the COP defect result, as detailed below:

[0090] In some embodiments, determining the COP defect result of the wafer under test based on the first score value, the second score value, and the third score value includes:

[0091] A comprehensive score is calculated based on the first score, the second score, the third score, the first weight parameter corresponding to the first score, the second weight parameter corresponding to the second score, and the third weight parameter corresponding to the third score.

[0092] If the overall score is greater than or equal to a preset score threshold, it is determined that a COP defect exists in the wafer under test.

[0093] In this embodiment, the calculation of the comprehensive score (denoted as Score) can be expressed as:

[0094] Score = A·S1+ B·S2 +C·S3;

[0095] Wherein, S1 is the first score, S2 is the second score, S3 is the third score, A is the first weight parameter, B is the second weight parameter, and C is the third weight parameter.

[0096] For example, in the initial stage, weight parameters A, B, and C can be set based on engineering experience. During system operation, the system compares the actual re-examination results from scanning electron microscopy (SEM) with the predicted results. If misjudgments or omissions occur, the system automatically fine-tunes the weight parameters, causing the scoring results of subsequent similar samples to converge in the correct direction. In this way, through the above feedback mechanism, the determination of the comprehensive score can gradually optimize the influence weights of information in each dimension without changing the overall judgment logic. This reduces the over-detection rate and improves the stability of COP recognition in long-term operation. Through continuous data accumulation, long-term stability and a decrease in the over-detection rate of single-piece COP recognition are achieved.

[0097] In this embodiment of the invention, the wafer under test is scanned and inspected to obtain the multi-channel light scattering feature vector of each defect point on the wafer under test and the basic detection results; then, based on the multi-channel light scattering feature vector, the probability of COP of a single defect point is initially judged at the optical response level, but it is not directly used as the COP detection result of the wafer under test; further, the wafer under test is comprehensively scored from the dimensions of the multi-channel light scattering feature vector, the dimensions of the global defect type composition, and the dimensions of the spatial distribution of each defect point, in combination with the proportion of the first type of defect points on the wafer under test and the spatial distribution of each defect point, and finally the COP defect detection result is determined, so as to achieve accurate identification of COP defects of a single wafer and improve the reliability of the detection results.

[0098] This invention also provides a COP defect detection device, such as... Figure 3 As shown, it includes:

[0099] The detection module 31 is used to scan and detect the wafer under test, and obtain the detection results of each defect point on the wafer under test. The detection results include the multi-channel light scattering feature vector, position information and optically determined defect type of each defect point. The multi-channel light scattering feature vector is used to represent the optical scattering characteristics of the defect point in different detection channels. The optically determined defect type includes the type of adhesion defect and the type of non-adhesion defect.

[0100] The first determining module 32 is used to determine a first score value for a first defect point as a crystal native particle COP defect based on the multi-channel light scattering feature vector. The first defect point is any defect point on the wafer to be tested. The first score value is used to quantify the probability that the first defect point is a COP defect based on the dimension of the multi-channel light scattering feature of the defect point.

[0101] The second determining module 33 is used to determine a second score value for the presence of COP defects in the wafer under test based on the proportion of the number of first type defect points on the wafer under test. The first type is the non-adhesive defect type indicated in the optically determined defect type. The second score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the dimension of the proportion of defect types corresponding to defect points.

[0102] The third determining module 34 is used to determine a third score value for the presence of COP defects in the wafer under test based on the spatial distribution of each defect point on the wafer under test. The spatial distribution is determined according to the position information of each defect point in the detection result. The third score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the dimension of the spatial distribution of the defect points.

[0103] The fourth determining module 35 is used to determine the COP defect result of the wafer under test based on the first score value, the second score value and the third score value.

[0104] In some embodiments, the first determining module 32 is specifically used for:

[0105] Obtain the multi-channel light scattering feature vector of the sample defect point, wherein the sample defect point is a predetermined non-adhesive defect point;

[0106] Calculate the first similarity between the multi-channel light scattering feature vector of the first defect point and the multi-channel light scattering feature vector of the sample defect point;

[0107] The first similarity is converted into a score value to obtain a first score value, which is proportional to the first similarity.

[0108] In some embodiments, the second determining module 33 is specifically used for:

[0109] The defect type of each defect point in the detection results is analyzed to obtain the number of defect points of the first type and the number of defect points of the second type, and the adhesion defect type indicated in the optically determined defect type of the second type.

[0110] The proportion of the first type of defect points is determined based on the number of the first type of defect points and the number of the second type of defect points;

[0111] The percentage of defects of the first type is converted into a score value to obtain the second score value, which is proportional to the percentage of defects of the first type.

[0112] In some embodiments, the third determining module 34 is specifically used for:

[0113] Based on the location information of each defect point in the detection results, the target spatial distribution characteristics of each defect point on the wafer under test are determined;

[0114] Calculate the second similarity between the target spatial distribution feature and the first spatial distribution feature, the third similarity between the target spatial distribution feature and the second spatial distribution feature, and the fourth similarity between the target spatial distribution feature and the third spatial distribution feature. The first spatial distribution feature indicates that the defect points only gather in the central region of the wafer, the second spatial distribution feature indicates that the defect points only gather in the edge region of the wafer, and the third spatial distribution feature indicates that the defect points gather in both the central region and the edge region of the wafer.

[0115] The fifth similarity is converted into a score value to obtain a third score value, which is proportional to the fifth similarity. The fifth similarity is the maximum value among the second, third, and fourth similarities.

[0116] In some embodiments, the fourth determining module 35 is specifically used for:

[0117] A comprehensive score is calculated based on the first score, the second score, the third score, the first weight parameter corresponding to the first score, the second weight parameter corresponding to the second score, and the third weight parameter corresponding to the third score.

[0118] If the overall score is greater than or equal to a preset score threshold, it is determined that a COP defect exists in the wafer under test.

[0119] In some embodiments, the detection module 31 is specifically used for:

[0120] The wafer under test is scanned and inspected using a particle counter to obtain the Klarf file output by the particle counter and the multi-channel light scattering signal of each defect point on the wafer under test in different detection channels. The detection channels include a normal incidence channel, an oblique incidence channel and a bright field detection channel.

[0121] The Klarf file is parsed to obtain the location information of each defect point on the wafer under test and the defect type determined by optical analysis;

[0122] Feature extraction is performed on the multi-channel light scattering signal to obtain the multi-channel light scattering feature vector.

[0123] It should be understood that the COP defect detection device is capable of implementing each process of the above-described method, with each technical feature corresponding to the other, and achieving the same technical effect. To avoid repetition, it will not be described again here.

[0124] Please refer to Figure 4 The present invention also provides an electronic device 40, including a processor 41, a memory 42, and a computer program stored in the memory 42 and executable on the processor 41. When the computer program is executed by the processor 41, it implements the various processes of the above-described COP defect detection method embodiments and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0125] This invention also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described COP defect detection method embodiments and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0126] The computer-readable storage medium, as defined herein, includes both permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage, or any other non-transfer medium that can be used to store information accessible to the terminal device under test. As defined herein, computer-readable storage media does not include transient media, such as modulated data signals and carrier waves.

[0127] This invention also provides a computer program product, including computer instructions that, when executed by a processor, implement the above-described... Figure 1 The various processes of the method embodiments shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0128] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0129] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0130] In the various method embodiments of this disclosure, the sequence numbers of each step are not intended to limit the order of the steps. For those skilled in the art, any changes in the order of the steps are within the scope of protection of this disclosure without any creative effort.

[0131] It should be noted that the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, since the embodiments are basically similar to the product embodiments, the descriptions are relatively simple, and the relevant parts can be referred to the descriptions of the product embodiments.

[0132] The above description represents the preferred embodiments of this disclosure. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles described herein, and these improvements and modifications should also be considered within the scope of protection of this disclosure.

Claims

1. A method for detecting COP defects, characterized in that, include: The wafer under test is scanned and inspected to obtain the detection results of each defect point on the wafer under test. The detection results include the multi-channel light scattering feature vector, position information and optically determined defect type of each defect point. The multi-channel light scattering feature vector is used to represent the optical scattering characteristics of the defect point in different detection channels. The optically determined defect type includes adhesive defect type and non-adhesive defect type. Based on the multi-channel light scattering feature vector, a first score value is determined for the first defect point to be a crystal native particle COP defect. The first defect point is any defect point on the wafer under test. The first score value is used to quantify the probability that the first defect point is a COP defect based on the dimension of the multi-channel light scattering feature of the defect point. Based on the proportion of the first type of defect points on the wafer under test, a second score value is determined for the presence of COP defects in the wafer under test. The first type is the non-adhesive defect type indicated in the optically determined defect type. The second score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the proportion of defect points corresponding to the defect type. Based on the spatial distribution of each defect point on the wafer under test, a third score value is determined for the presence of COP defects in the wafer under test. The spatial distribution is determined according to the position information of each defect point in the detection results. The third score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the dimension of the spatial distribution of the defect points. The COP defect result of the wafer under test is determined based on the first score, the second score, and the third score.

2. The method according to claim 1, characterized in that, The determination of the first defect point as a crystal native particle COP defect based on the multi-channel light scattering feature vector includes: Obtain the multi-channel light scattering feature vector of the sample defect point, wherein the sample defect point is a predetermined non-adhesive defect point; Calculate the first similarity between the multi-channel light scattering feature vector of the first defect point and the multi-channel light scattering feature vector of the sample defect point; The first similarity is converted into a score value to obtain a first score value, which is proportional to the first similarity.

3. The method according to claim 1, characterized in that, The determination of a second score value for the presence of COP defects in the wafer under test based on the proportion of first-type defects on the wafer under test includes: The defect type of each defect point in the detection results is analyzed to obtain the number of defect points of the first type and the number of defect points of the second type, and the adhesion defect type indicated in the optically determined defect type of the second type. The proportion of the first type of defect points is determined based on the number of the first type of defect points and the number of the second type of defect points; The percentage of defects of the first type is converted into a score value to obtain the second score value, which is proportional to the percentage of defects of the first type.

4. The method according to claim 1, characterized in that, The third score for determining the presence of COP defects in the wafer under test based on the spatial distribution of each defect point on the wafer under test includes: Based on the location information of each defect point in the detection results, the target spatial distribution characteristics of each defect point on the wafer under test are determined; Calculate the second similarity between the target spatial distribution feature and the first spatial distribution feature, the third similarity between the target spatial distribution feature and the second spatial distribution feature, and the fourth similarity between the target spatial distribution feature and the third spatial distribution feature. The first spatial distribution feature indicates that the defect points only gather in the central region of the wafer, the second spatial distribution feature indicates that the defect points only gather in the edge region of the wafer, and the third spatial distribution feature indicates that the defect points gather in both the central region and the edge region of the wafer. The fifth similarity is converted into a score value to obtain a third score value, which is proportional to the fifth similarity. The fifth similarity is the maximum value among the second, third, and fourth similarities.

5. The method according to any one of claims 1 to 4, characterized in that, The step of determining the COP defect result of the wafer under test based on the first score, the second score, and the third score includes: A comprehensive score is calculated based on the first score, the second score, the third score, the first weight parameter corresponding to the first score, the second weight parameter corresponding to the second score, and the third weight parameter corresponding to the third score. If the overall score is greater than or equal to a preset score threshold, it is determined that a COP defect exists in the wafer under test.

6. The method according to any one of claims 1 to 4, characterized in that, The scanning and inspection of the wafer under test, to obtain the detection results of each defect point on the wafer under test, includes: The wafer under test is scanned and detected using a particle counter to obtain the Klarf file output by the particle counter and the multi-channel light scattering signal of each defect point on the wafer under test in different detection channels. The detection channels include a normal incidence channel, an oblique incidence channel and a bright field detection channel. The Klarf file is parsed to obtain the location information of each defect point on the wafer under test and the defect type determined by optical analysis; Feature extraction is performed on the multi-channel light scattering signal to obtain the multi-channel light scattering feature vector.

7. A COP defect detection device, characterized in that, include: The detection module is used to scan and detect the wafer under test, and obtain the detection results of each defect point on the wafer under test. The detection results include the multi-channel light scattering feature vector, position information and optically determined defect type of each defect point. The multi-channel light scattering feature vector is used to represent the optical scattering characteristics of the defect point in different detection channels. The optically determined defect type includes adhesive defect type and non-adhesive defect type. The first determining module is used to determine a first score value for a first defect point as a crystal native particle COP defect based on the multi-channel light scattering feature vector. The first defect point is any defect point on the wafer under test. The first score value is used to quantify the probability that the first defect point is a COP defect based on the dimension of the multi-channel light scattering feature of the defect point. The second determining module is used to determine a second score value for the presence of COP defects in the wafer under test based on the proportion of the number of first type defect points on the wafer under test. The first type is the non-adhesive defect type indicated in the optically determined defect type. The second score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the dimension of the proportion of defect points corresponding to the defect type. The third determining module is used to determine a third score value for the presence of COP defects in the wafer under test based on the spatial distribution of each defect point on the wafer under test. The spatial distribution is determined according to the position information of each defect point in the detection result. The third score value is used to quantify the probability of the presence of COP defects in the wafer under test based on the dimension of the spatial distribution of the defect points. The fourth determining module is used to determine the COP defect result of the wafer under test based on the first score value, the second score value and the third score value.

8. An electronic device, characterized in that, include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, Includes computer instructions that, when executed by a processor, implement the steps of the method as described in any one of claims 1 to 6.

Images