Reliability determination method for wafer defect detection system

Abstract

The present application provides a reliability determination method for a wafer defect detection system, comprising: determining standard detection data for defective wafers in mass production; performing defect detection on the defective wafers using the detection system to be analyzed to obtain sample detection data; and determining the reliability of the detection system to be analyzed based on the sample detection data and the standard detection data. The standard defect data in the reliability determination method provided by the present application is obtained based on defective wafers in actual mass production, and the defects of the actual mass-produced wafers can better reflect real wafer defects, making the standard detection data more authentic and the reliability assessment of the detection system to be analyzed more accurate.

Description

Technical Field

[0001] The present application relates to the technical field of reliability analysis of detection systems, and more specifically, to a reliability determination method for a wafer defect detection system. Background Art

[0002] The semiconductor manufacturing process is typically divided into front-end (FOL) and back-end (BOL) processes. The FOL mainly includes wafer fabrication (Fab) and wafer probing, while the BOL mainly includes IC packaging and testing processes including initial testing and final testing.

[0003] In the wafer fab process, defect scanning machines are frequently used, and their results are often used as an important basis for problem analysis and yield improvement. However, there is no systematic and scientific solution to measure the reliability of their detection results. Summary of the Invention

[0004] The present application provides a reliability determination method for a wafer defect detection system that can at least partially solve the above-mentioned problems existing in the related art.

[0005] The present application provides a method for determining the reliability of a wafer defect detection system, comprising: determining standard detection data of defective wafers in mass production; detecting the defective wafers using the detection system to be analyzed to obtain sample detection data; and determining the reliability of the detection system to be analyzed based on the sample detection data and the standard detection data.

[0006] In some embodiments, determining standard inspection data for defective wafers in mass production includes: performing defect inspection on the defective wafers using a standard inspection system to obtain an original inspection database; and determining a plurality of defect data and a plurality of non-defect data from the original inspection database, wherein the determined plurality of defect data and the plurality of non-defect data constitute standard inspection data, wherein the number of the plurality of defect data is greater than the number of the plurality of non-defect data.

[0007] In some embodiments, the defect data includes the relative coordinates of the defect position in the defective wafer and the identification of the defect; the non-defect data includes the relative coordinates of the non-defect position in the defective wafer and the identification of the non-defect.

[0008] In some embodiments, the standard inspection system includes a scanning machine, and the original inspection database includes a plurality of defect images obtained by performing multiple defect scans on a defective wafer using the scanning machine; wherein, determining a plurality of defect data and a plurality of non-defect data from the original inspection database includes: stacking the plurality of defect images to determine: a plurality of defect positions having the same relative coordinates and defect identification in each defect image; and a plurality of non-defect positions having the same relative coordinates and non-defect identification in each defect image; wherein the relative coordinates of the plurality of defect positions and the defect identification constitute the defect data; and the relative coordinates of the plurality of non-defect positions and the non-defect identification constitute the non-defect data.

[0009] In some embodiments, the defective wafer includes a central region and a peripheral region; a plurality of defective locations and a plurality of non-defective locations are dispersed in the central region and the peripheral region.

[0010] In some embodiments, the surface of the defective wafer includes multiple partitions, multiple defect locations and multiple non-defect locations are located in the multiple partitions, one partition is located in the central area, and the other partitions are distributed in the peripheral area.

[0011] In some implementations, the sum of the number of defective data and the number of non-defective data is greater than or equal to 30.

[0012] In some embodiments, a defective wafer is inspected using the inspection system to be analyzed to obtain sample inspection data, including: inspecting the defective wafer using the inspection system to be analyzed to obtain inspection results for multiple defective positions and multiple non-defective positions of the defective wafer; identifying the inspection results at the multiple defective positions; and identifying the inspection results at the multiple non-defective positions.

[0013] In some embodiments, the sample detection data includes multiple sub-sample detection databases, and the multiple sub-sample detection databases are for detecting defective wafers using multiple detection systems to be analyzed, respectively obtaining detection results at multiple defect locations and multiple non-defect locations of the defective wafer, and performing defect identification and non-defect identification on the detection results.

[0014] In some embodiments, each sub-sample detection database includes multiple defect data, and the multiple defect data are for each detection system to be analyzed to perform multiple detections on the defective wafer, respectively obtaining detection results at multiple defect locations and multiple non-defect locations of the defective wafer, and performing defect identification and non-defect identification on the detection results.

[0015] In some embodiments, the reliability of the detection system to be analyzed is determined based on the sample detection data and the standard detection data, including: determining the consistency and validity of the detection system to be analyzed based on the defect identification and non-defect identification in the sample detection data and the defect identification and non-defect identification in the standard detection data; and determining the reliability of the detection system to be analyzed based on the consistency and validity of the detection system to be analyzed.

[0016] In the reliability determination method of the wafer defect detection system provided in the embodiment of the present application, the standard defect data is obtained based on defective wafers actually produced in mass production, and the defects of the wafers actually produced in mass production can better reflect the real wafer defects, so that the authenticity of the standard detection data is higher and the reliability evaluation of the detection system to be analyzed is more accurate.

[0017] In the reliability determination method of the wafer defect detection system provided in the embodiment of the present application, when determining the standard test data, in addition to using the defect data of bad wafer particles (bad die), a certain proportion of non-defect data of good wafer particles is also added, so that when determining the reliability of the detection system to be analyzed, in addition to referring to the ability of defect data to be accurately discovered, the ability of non-defect data to be accurately discovered is also referred to. Therefore, when performing the reliability analysis of the detection system to be analyzed, the present application takes into account both the ability to identify defect data and the ability to misjudge non-defect data, and makes judgments from the two perspectives of loss and overkill, which meets the relevant requirements of the industry and is closer to the actual situation.

[0018] In the reliability determination method of the wafer defect detection system provided in the embodiment of the present application, all bad dies and good dies are judged based on the consistency of coordinate positions, which can improve the accuracy of the judgment of defect data and non-defect data, so that in subsequent steps, the detection system to be analyzed can accurately determine its ability to recognize defect data and misjudge non-defect data. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Other features, objects and advantages of the present application will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings, in which:

[0020] Figure 1A It is the defect data obtained by scanning a standard film using a golden tool in the related art;

[0021] Figure 1B It is the defect data after the first scan of the standard film using the scanning machine (study tool) to be analyzed in the related art;

[0022] Figure 1CIt is the defect data after the second scan of the standard film using the scanning machine (study tool) to be analyzed in the related art;

[0023] Figure 2 is a flow chart of a reliability determination method 1000 of a wafer defect detection system according to an embodiment of the present application;

[0024] Figure 3 According to the embodiment of this application Figure 2 Flowchart of S110;

[0025] Figure 4 is a schematic diagram of the distribution of sampling positions on a wafer in standard test data according to an embodiment of the present application;

[0026] Figure 5 According to the embodiment of this application Figure 2 Flowchart of S120;

[0027] Figure 6 According to the embodiment of this application Figure 2 Flowchart of S130;

[0028] Figure 7 is a schematic diagram of a consistency report of a detection system according to an embodiment of the present application; and

[0029] Figure 8 It is a schematic diagram of the effectiveness report of the detection system according to the implementation method of the present application. DETAILED DESCRIPTION

[0030] For a better understanding of the present application, various aspects of the present application will be described in more detail with reference to the accompanying drawings. It should be understood that these detailed descriptions are merely descriptions of exemplary embodiments of the present application and are not intended to limit the scope of the present application in any way. Throughout the specification, the same reference numerals refer to the same elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0031] It should be noted that in this specification, the expressions first, second, third, etc. are only used to distinguish one feature from another feature area, and do not represent any limitation on the features, and especially do not represent any order of precedence.

[0032] As used herein, the terms "substantially," "about," and similar terms are used as terms of approximation, not degree, and are intended to account for the inherent variations in measured or calculated values ​​that would be recognized by one of ordinary skill in the art.

[0033] It should also be understood that expressions such as "comprises," "including," "having," "includes," and / or "comprising" are open rather than closed expressions in this specification, indicating the presence of the stated features, elements, and / or components, but do not exclude the presence of one or more other features, elements, components, and / or combinations thereof. In addition, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features rather than just the individual elements in the list. In addition, when describing embodiments of the present application, "may" is used to mean "one or more embodiments of the present application." And, the term "exemplary" is intended to refer to an example or illustration.

[0034] Unless otherwise defined, all words used herein (including engineering terms and scientific and technological terms) have the same meaning as commonly understood by those skilled in the art to which this application belongs. It should also be understood that, unless otherwise specified in this application, words defined in commonly used dictionaries should be interpreted as having the same meaning as they do in the context of the relevant technology, and should not be interpreted in an idealized or overly formal sense.

[0035] It should be noted that, unless otherwise specified or inconsistent with the context, the embodiments and features of the embodiments in this application may be combined with each other. Furthermore, unless expressly limited or inconsistent with the context, the specific steps included in the methods described in this application are not necessarily limited to the order in which they are described, but may be performed in any order or in parallel. The present application will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0036] Regarding reliability analysis of defect scanning machines, there are two common approaches in related technologies.

[0037] Method 1: Use the manufacturer's standard film scan results to establish a reference value for monitoring the scanner's detection accuracy and stability. The standard film scan results are obtained by scanning the standard film with the manufacturer's first released scanner, serving as the reference standard.

[0038] For example, ten scanners (sample numbers 1 to 10) were used to scan the manufacturer's standard film three times. The number of defects (defectives) obtained in each scan was recorded, and the ratio of the defect count to the reference value was calculated to obtain the sampling efficiency (capture rate). The scan results were obtained according to the following rules, as shown in Table 1, where "1" indicates a reliable result and "0" indicates an unreliable result. Table 2 shows the cross-analysis table (MSA Kappa) based on the data in Table 1.

[0039] When the reference value is ≥30; capture rate ≥85% is judged as "1"; capture rate <85% is judged as "0".

[0040] When the reference value is less than 30, if the difference between the test result and the reference value is ≤5, it is judged as "1"; if the difference between the test result and the reference value is greater than 5, it is judged as "0".

[0041] Table 1: Sampling efficiency (capture rate) data record table

[0042]

[0043] Table 2: Cross-analysis table (MSA Kappa)

[0044]

[0045] Kappa < 0.40, poor consistency between the evaluator and the reference

[0046] Kappa>0.75, the consistency between the evaluator and the reference is very good

[0047] Kappa = 1, perfect agreement between the evaluator and the reference

[0048] The Kappa table shows that there is perfect agreement between the evaluator and the reference.

[0049] As can be seen in Table 1, the calculations only consider the defect locations on the wafer, not the accuracy of identifying completely defect-free locations. Furthermore, all capture rates are typically determined to be 1, and the Kappa index is also calculated to be 1 (see Table 2). This clearly does not meet the requirement for counting-based inspection systems to analyze samples with a certain percentage of both out-of-specification and in-specification samples.

[0050] Moreover, the theoretical support for the above rules is insufficient for judging 1 if the capture rate exceeds 85% when the reference value exceeds 30 and judging 1 if the difference in the number of defects does not exceed 5 when the reference value is less than 30.

[0051] In addition, the above method has the following problems:

[0052] 1. Whether the scanning equipment debugged by the original manufacturer can accurately detect the actual situation of wafer defects and whether it is always stable. These factors are uncertain.

[0053] 2. After the scanner is released, a fixed-frequency standard film scanning monitoring method is used. If the number of particles on the wafer is found to be excessive and exceeds the control limit (Out of Control; OOC), it is difficult to accurately determine whether the cause is due to changes in the standard film or changes in the scanner's recognition capabilities.

[0054] 3. Using the first released scanner as a reference standard to specify the relative capture rate of subsequent scanners is not a viable approach to system analysis that meets industry requirements.

[0055] 4. The capture rate calculation method only identifies the number of wafer defects and ignores the accuracy differences when the same number of wafer defects is detected at different locations.

[0056] Method 2: Directly use the standard film and perform multiple tests with the scanning machine to be analyzed to perform counting detection system analysis.

[0057] For example, a standard wafer is used as a reference wafer, and a golden tool is used to scan the reference wafer once and record defect data, such as Figure 1A As shown, a total of 24 defect data from No. 1 to No. 24 are recorded. Then, the reference wafer is scanned several times, for example, twice, using the scanning tool to be analyzed, and the defect data are recorded respectively. The capture rate is calculated, as shown in FIG. Figure 1B and Figure 1C shown.

[0058] analyze Figure 1B It can be seen that the scanning machine to be analyzed samples 24 defect locations after scanning the standard film, of which 22 defect locations are Figure 1A The 22 defect locations shown in the figure are the same, and 2 defect locations are the same as Figure 1A The two defects shown in the figure are located in different locations. This shows that although the capture rate of defects in this scanning process was 100% (24 / 24), defects originally detected by the golden tool were missed, while "defects" in new locations were also detected. This omission and addition of the same number of defects masked the actual inaccuracy of the scanning tool under analysis.

[0059] analyze Figure 1C It can be seen that the scanning machine to be analyzed samples 26 defect locations after scanning the standard film, of which 24 defect locations are Figure 1A The 24 defect locations shown in the figure are the same, and there are 2 defect locations. Figure 1A The defect locations are not shown in the figure. It can be seen that the capture rate of the number of defects in this scanning process has exceeded 100% (26 / 24). Moreover, whether the two newly added "defects" detected by the scanning machine to be analyzed are real defects remains to be further confirmed.

[0060] In addition, this method has the following problems: the wafer defects of standard wafers are different from the wafer defects of actual mass-produced wafers. It is easier to identify wafer defects of standard wafers than wafer defects of actual wafers. Therefore, the detection system analysis based on standard wafers has little practical significance.

[0061] Based on this, an embodiment of the present application provides a method for determining the reliability of a wafer defect detection system. Figure 2 A flowchart of a method 1000 for determining reliability of a wafer defect detection system according to an embodiment of the present application is shown, comprising the following steps:

[0062] S110, determining standard inspection data of defective wafers 100 in mass production;

[0063] S120 , using the inspection system to be analyzed to inspect the defective wafer 100 to obtain sample inspection data; and

[0064] S130: Determine the reliability of the detection system to be analyzed based on the sample detection data and the standard detection data.

[0065] It should be understood that the steps shown in reliability determination method 1000 are not exclusive, and other steps may be performed before, after, or between any of the steps shown. In addition, some of the steps may be performed simultaneously or in different steps. Figure 2 Executed in the order shown.

[0066] Figure 3 、 Figure 5 and Figure 6 Shown respectively Figure 2 Flowchart of step S110, step S120 and step S130, Figure 4 FIG1 shows a schematic diagram of the distribution of sampling positions on a wafer in standard detection data according to an embodiment of the present application. Figure 7 and Figure 8 The following diagrams respectively show the consistency and effectiveness of the detection system according to the embodiment of the present application. Figures 3 to 8 The above steps S110 to S130 are further described.

[0067] S110 , determining standard inspection data of defective wafers 100 in mass production.

[0068] The defective wafer 100 in step S110 may be an actual mass-produced wafer. The defects of the actual mass-produced wafer can better reflect the real wafer defects, making the authenticity of the standard detection data higher and the reliability evaluation of the detection system to be analyzed more accurate.

[0069] In some embodiments, as Figure 3 As shown, determining the standard inspection data of the defective wafer 100 in mass production in step S110 includes the following steps:

[0070] S112, performing defect detection on the defective wafer 100 using a standard detection system to obtain an original detection database; and

[0071] S114, determining a plurality of defect data and a plurality of non-defect data from the original detection database, wherein the determined plurality of defect data and the plurality of non-defect data constitute standard detection data, wherein the number of the plurality of defect data is greater than the number of the plurality of non-defect data.

[0072] Exemplarily, I defective data and J non-defective data can be determined from the original test database, where I is greater than J. These I defective data and J non-defective data constitute the standard test data in step S110. Optionally, to ensure a sufficient number of samples, the sum of the number of defective data and the number of non-defective data is greater than or equal to 30, that is, the sum of I and J is greater than or equal to 30. Exemplarily, the present application can determine 40 defective data and 10 non-defective data to constitute the standard test data.

[0073] In some embodiments, the standard inspection system in step S112 may be the manufacturer's earliest released inspection system, which has industry-recognized reliability. In other words, by using the standard inspection system to perform defect inspection on defective wafers 100, thereby obtaining an original inspection database, the defect data in the original inspection database has a high degree of credibility.

[0074] It can be understood that the defect data is data at the location of a bad die on the wafer, and the non-defect data is data at the location of a good die on the wafer.

[0075] In the above scheme, in the process of determining the standard test data, in addition to using defect data of bad wafer particles, a certain proportion of non-defective data of good wafer particles is also added. Therefore, when determining the reliability of the test system to be analyzed, in addition to the ability to accurately detect defect data, the ability to accurately detect non-defect data is also referenced. Therefore, when performing the reliability analysis of the test system to be analyzed, this application takes into account both the ability to identify defect data and the ability to misjudge non-defect data, and judges from the two perspectives of loss and overkill, which meets the relevant requirements of the industry and is closer to the actual situation.

[0076] In some embodiments, the defect data includes the relative coordinates of the defect position in the defective wafer 100 and the identification of the defect; the non-defect data includes the relative coordinates of the non-defect position in the defective wafer 100 and the identification of the non-defect.

[0077] Optionally, a rectangular coordinate system with the center as the origin is established on the surface of the defective wafer 100. When I defect data is determined, the relative coordinates of each defect location on the defective wafer 100 are simultaneously obtained and the defects are marked for defect comparison in subsequent steps. Similarly, when J non-defect data are determined, the relative coordinates of each non-defect location on the defective wafer 100 are simultaneously obtained and the non-defects are marked for non-defect comparison in subsequent steps.

[0078] For example, the defect identifier may be “1” and the non-defect identifier may be “0.” It is understandable that the defect identifier and the non-defect identifier may also be other letters and / or numbers, which are not limited in this application.

[0079] In the above scheme, the coordinates of each defective position and non-defective position in the defective wafer 100 are obtained, that is, all bad dies and good dies are judged based on the consistency of coordinate positions, which can improve the accuracy of judging defective data and non-defective data, so that in subsequent steps, the detection system to be analyzed can accurately determine its ability to recognize defective data and its ability to misjudge non-defective data.

[0080] In some embodiments, the inspection system includes a scanning machine, and the original inspection database includes a plurality of defect images obtained by performing multiple defect scans on the defective wafer 100 using the scanning machine.

[0081] In step S114, determining multiple defect data and multiple non-defect data from the original inspection database includes stacking multiple defect images to determine: multiple defect locations having the same relative coordinates and defect identifiers in each defect image; and multiple non-defect locations having the same relative coordinates and non-defect identifiers in each defect image. The relative coordinates of the multiple defect locations and the defect identifiers constitute the defect data, and the relative coordinates of the multiple non-defect locations and the non-defect identifiers constitute the non-defect data.

[0082] Exemplarily, a scanner is used to perform M defect scans on the defective wafer 100 and obtain M defect images. Optionally, to improve the credibility of the standard test data, M is an integer greater than or equal to 3. Optionally, M in this application is 5. That is, the scanner is used to scan the defective wafer 100 5 times, obtaining 5 defect images of the defective wafer 100.

[0083] Then, after stacking the five defect images, 50 locations with consistent detection results were identified, including 40 locations with consistent detection results (marked as "1") and 10 locations with consistent detection results (marked as "0").

[0084] In the above scheme, when determining the standard inspection data, the defective wafer 100 is scanned multiple times, and then the defect patterns are stacked to determine the position where the multiple scanning results are defects as the reference defect point, and the position where the multiple scanning results are non-defective is determined as the reference non-defect point, thereby improving the accuracy of the standard inspection data.

[0085] In some embodiments, the defective wafer 100 includes a central region 102 (center) and an outer region 104 (edge); a plurality of defect locations and a plurality of non-defect locations are dispersed within the central region 102 and the outer region 104 .

[0086] In the above scheme, when determining the standard test data, the defective positions (bad spots) and non-defective positions (no bad spots) of the defective wafer 100 selected should cover as many areas of the defective wafer 100 (cover wafer) as possible, thereby increasing the randomness of sampling and thus increasing the randomness of sampling of the sample test data (refer to step S120), which is beneficial to improving the accuracy of determining the reliability of the test system to be analyzed.

[0087] In some embodiments, the surface of the defective wafer 100 includes a plurality of partitions 106 , wherein a plurality of defect locations and a plurality of non-defect locations are located in the plurality of partitions 106 . One partition 106 is located in the central region 102 , and the other partitions 106 are evenly distributed in the peripheral region 104 .

[0088] Exemplarily, the surface of the defective wafer 100 includes S partitions 106 , wherein one partition 106 is located in the central region 102 , and S−1 partitions 106 are evenly distributed in the peripheral region 104 .

[0089] Optionally, S is 5, that is, the surface of the defective wafer 100 includes 5 partitions 106, namely the first region, the second region ... the fifth region, numbered 1 to 5, referring to Figure 4 Then, the defective positions of I defective data and the non-defective positions of J non-defective data are dispersedly located in the areas numbered 1 to 5.

[0090] Furthermore, the first through fifth regions can substantially cover all directions of the defective wafer 100. For example, the first region numbered 1 is located in the central region 102 of the defective wafer 100, and the second, third, fourth, and fifth regions numbered 2 through 5 are located in the peripheral region 104 of the defective wafer 100. In some embodiments, the second, third, fourth, and fifth regions numbered 2 through 5 are all located in the peripheral region 104 of the defective wafer 100. For example, the areas of the first through fifth regions can be the same or different, and this application does not limit this.

[0091] In the above scheme, when determining standard inspection data, defective wafers 100 are partitioned and classified as either defective or non-defective. This allows the reliability of the inspection system to be analyzed, with its ability to identify defects in the center and edge, as well as its ability to misjudge non-defects, to be analyzed separately. Furthermore, when the final inspection result is a non-defective (NG), the cause is more easily analyzed.

[0092] In some embodiments of the present application, standard test data can be referred to as shown in Table 3.

[0093] Table 3: Standard test data table

[0094]

[0095] In Table 3, the area number is the number of the partition 106 where the defect position is located on the defective wafer 100, the defect number / non-defect number is the serial number of the defective data or non-defective data, the coordinates are the relative coordinates of the defective position or non-defective position on the defective wafer 100, the result identifier "1" represents that the recognition result is defective data, and the result identifier "0" represents that the recognition result is non-defective data.

[0096] S120 , using the inspection system to be analyzed to inspect the defective wafer 100 to obtain sample inspection data.

[0097] In some embodiments, the sample inspection data is obtained by inspecting the defective wafer 100 using the inspection system to be analyzed, obtaining inspection results of the defective wafer 100 at multiple defective locations and multiple non-defective locations, and marking the inspection results.

[0098] In some embodiments, as Figure 5 As shown, in step S120, the defective wafer 100 is inspected by the inspection system to be analyzed to obtain sample inspection data, including:

[0099] S122, using the inspection system to be analyzed to inspect the defective wafer, and obtaining inspection results of multiple defective locations and multiple non-defective locations of the defective wafer; and

[0100] S124, identifying the detection results at multiple defect locations and the detection results at multiple non-defect locations.

[0101] It should be noted that the relative coordinates of the multiple defect positions on the defective wafer obtained using the inspection system to be analyzed in step S122 are the same as the relative coordinates of the multiple defect positions on the defective wafer obtained using the standard inspection system in step S110. The relative coordinates of the multiple non-defect positions on the defective wafer obtained using the inspection system to be analyzed in step S122 are the same as the relative coordinates of the multiple non-defect positions on the defective wafer obtained using the standard inspection system in step S110.

[0102] Furthermore, in step S124, when marking the inspection results at multiple defect locations, the same phenomenon at the same location can be determined as a defective item and identified, and marked as "1". Different locations or different phenomena are both determined as defective items and not identified, and marked as "0". When marking the inspection results at multiple non-defective locations, the same phenomenon at the same location can be determined as no defective item and identified, and marked as "0". Different locations or different phenomena are both determined as no defective item and not identified, and marked as "1".

[0103] In some embodiments, the sample detection data includes multiple sub-sample detection databases, and the multiple sub-sample detection databases are for detecting the defective wafer 100 using multiple detection systems to be analyzed, respectively obtaining detection results at multiple defect positions and multiple non-defect positions of the defective wafer 100, and performing defect identification and non-defect identification on the detection results.

[0104] In the above solution, the sample test data is obtained by testing the defective wafer 100 using multiple test systems to be analyzed, and then the consistency between the multiple test systems to be analyzed can be obtained based on the sample test data and the standard test data.

[0105] Exemplarily, the sample detection data includes N sub-sample detection databases, where N is an integer greater than 1.

[0106] In some embodiments, each sub-sample detection database includes multiple defect data, and the multiple defect data are for each detection system to be analyzed to perform multiple detections on the defective wafer 100, and obtain detection results at multiple defect locations and multiple non-defect locations of the defective wafer 100, and perform defect identification and non-defect identification on the detection results.

[0107] In the above scheme, each detection system to be analyzed is tested multiple times to obtain multiple sample data, which can improve the accuracy of the data.

[0108] Exemplarily, each sub-sample detection database includes n pieces of defect data, where n is an integer greater than 1.

[0109] In some embodiments, N and n are each any integer between 3 and 5. For example, both N and n can be 3. The present application performs reliability analysis on three inspection systems to be analyzed: a first inspection system to be analyzed (Tool 1), a second inspection system to be analyzed (Tool 2), and a third inspection system to be analyzed (Tool 3). Each inspection system to be analyzed performs three inspections on the defective wafer 100 to obtain sample inspection data.

[0110] In some embodiments of the present application, the summary of standard test data and sample test data can be referred to as shown in Table 4.

[0111]

[0112]

[0113] S130: Determine the reliability of the detection system to be analyzed based on the sample detection data and the standard detection data.

[0114] After determining the standard test data in step S110 and confirming the sample test data in step S120, the sample test data and the standard test data can be compared. Generally speaking, if the sample test data has a high degree of overlap with the standard test data, the reliability of the test system to be analyzed is high. Conversely, if the sample test data has a low degree of overlap with the standard test data, the reliability of the test system to be analyzed is low.

[0115] However, in general, the industry has certain requirements for reliability analysis of detection systems. Figure 6 As shown, step S130 includes the following steps:

[0116] S132, determining the consistency and effectiveness of the detection system to be analyzed based on the defect identifiers and non-defect identifiers in the sample detection data and the defect identifiers and non-defect identifiers in the standard detection data; and

[0117] S134 , determining the reliability of the detection system to be analyzed based on the consistency and effectiveness of the detection system to be analyzed.

[0118] like Figure 7 The following table shows the consistency report of the test system to be analyzed based on the sample test data and standard test data in Table 4. Figure 7 It can be seen that for the first detection system to be analyzed, the second detection system to be analyzed, and the third detection system to be analyzed, the Kappa index in the consistency comparison results are 0.8562, 0.8973, and 0.8446, respectively, and the Kappa index in the comparison results with the standard are 0.9573, 0.8996, and 0.9398, respectively. In the consistency results between the defective data and non-defective data categories, the Kappa index is 0.8655.

[0119] It can be seen that the Kappa index in the consistency report is a value greater than 0 and less than 1, therefore, it can meet the industry's requirements for detection system analysis.

[0120] like Figure 8 The following table shows the effectiveness report of the detection system to be analyzed based on the sample detection data and standard detection data in Table 4. Figure 8 It can be seen that for the first detection system to be analyzed, the second detection system to be analyzed, and the third detection system to be analyzed, the effectiveness indexes are 98.6667, 96.6667, and 98.0000, respectively.

[0121] also, Figure 8 The validity report also includes a qualification report. Analysis of the qualification report shows that for the first, second, and third detection systems to be analyzed, the false positive probabilities are 0.0000, 0.0333, and 0.0250, respectively, and the false negative probabilities are 0.0667, 0.0333, and 0.0000, respectively.

[0122] Based on the above analysis results, determine whether the capabilities of the detection system to be analyzed are acceptable. If not, further improvement plans are proposed and verified again.

[0123] This application achieves accurate measurement of the health of the detection system of the defect scanning machine by optimizing the defect sampling, defect recording method and analysis method in the analysis process of the defect scanning counting detection system, providing direction for the optimization of such detection systems, and at the same time providing reliable data support for process improvement, effectively reducing the risks that may be brought about by the detection system in large-scale automated production, and providing strong support for cost control and customer satisfaction.

[0124] The above description is merely an embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of protection involved in this application is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but should also cover other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the technical concept. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for determining the reliability of a wafer defect detection system, characterized in that: include: Determine the standard inspection data for defective wafers in mass production; Using the inspection system to be analyzed to inspect the defective wafer to obtain sample inspection data; as well as Determining the reliability of the detection system to be analyzed based on the sample detection data and the standard detection data, Among them, the standard inspection data for defective wafers in mass production includes: Performing defect detection on the defective wafer using a standard detection system to obtain an original detection database; and A plurality of defect data and a plurality of non-defect data are determined from the original detection database, and the determined plurality of defect data and a plurality of non-defect data constitute the standard detection data, wherein the number of the plurality of defect data is greater than the number of the plurality of non-defect data.

2. The reliability determination method according to claim 1, wherein: The defect data includes the relative coordinates of the defect position in the defective wafer and the identification of the defect; the non-defect data includes the relative coordinates of the non-defect position in the defective wafer and the identification of the non-defect.

3. The reliability determination method according to claim 2, wherein: The standard inspection system includes a scanning machine, and the original inspection database includes a plurality of defect images obtained by performing multiple defect scans on the defective wafer using the scanning machine; Wherein, determining a plurality of defect data and a plurality of non-defect data from the original detection database includes: A plurality of said defect images are stacked to determine: a plurality of defect locations having the same relative coordinates and defect identification in each of the defect images; and a plurality of non-defect locations having the same relative coordinates and non-defect identifiers in each of the defect images; The relative coordinates of the plurality of defect positions and the identifications of the defects constitute the defect data; the relative coordinates of the plurality of non-defect positions and the identifications of the non-defects constitute the non-defect data.

4. The reliability determination method according to claim 3, wherein: The defective wafer includes a central area and a peripheral area; The plurality of defect locations and the plurality of non-defect locations are dispersedly located in the central region and the peripheral region.

5. The reliability determination method according to claim 4, wherein: The surface of the defective wafer includes a plurality of partitions, the plurality of defect positions and the plurality of non-defect positions are located in the plurality of partitions, wherein one partition is located in the central area, and the other partitions are distributed in the peripheral area. The reliability determination method according to claim 1 , wherein: The sum of the number of the plurality of defective data and the number of the plurality of non-defective data is greater than or equal to 30.

7. The reliability determination method according to any one of claims 3 to 6, wherein: Detecting the defective wafer using the detection system to be analyzed to obtain sample detection data includes: Using the inspection system to be analyzed to inspect the defective wafer, and obtaining inspection results of the plurality of defective locations and the plurality of non-defective locations of the defective wafer; identifying detection results at the plurality of defect locations; and Inspection results at the plurality of non-defect locations are identified.

8. The reliability determination method according to claim 7, wherein: The sample detection data includes multiple sub-sample detection databases, and the multiple sub-sample detection databases are for detecting the defective wafer using multiple detection systems to be analyzed, respectively obtaining detection results at the multiple defect positions and the multiple non-defect positions of the defective wafer, and performing defect identification and non-defect identification on the detection results.

9. The reliability determination method according to claim 8, wherein: Each of the sub-sample detection databases includes multiple defect data, and the multiple defect data are for each detection system to be analyzed to perform multiple detections on the defective wafer, respectively obtain the detection results at the multiple defect positions and the multiple non-defect positions of the defective wafer, and perform defect identification and non-defect identification on the detection results.

10. The reliability determination method according to claim 8, wherein: Determining the reliability of the detection system to be analyzed based on the sample detection data and the standard detection data includes: Determining the consistency and effectiveness of the inspection system to be analyzed based on the defect identifiers and non-defect identifiers in the sample inspection data and the defect identifiers and non-defect identifiers in the standard inspection data; and The reliability of the detection system to be analyzed is determined based on the consistency and effectiveness of the detection system to be analyzed.

Images