Wafer defect sampling method and storage medium

Abstract

The invention discloses a wafer defect sampling method and a computer storage medium, and the method comprises the steps: obtaining a standard defect data file generated after a wafer defect detection machine scans a wafer, analyzing the standard defect data file, and extracting the coordinate of each defect and the feature parameter of the defect to form a defect information set; based on the coordinate of each defect and wafer area information, each defect is mapped to a specific area on the wafer, and the specific area on the wafer comprises a core area and a non-core area; respectively sampling defect information in the defect information set based on the types of the characteristic parameters, and preferentially collecting the defect information with high parameter values during sampling; and combining the collected information of all defects into a simplified defect list, and sending the simplified defect list to an SEM (scanning electron microscope) machine table for scanning imaging and defect review.

Description

Technical Field

[0001] This invention relates to the field of wafer testing technology, and in particular to a wafer defect sampling method and storage medium. Background Technology

[0002] In semiconductor manufacturing processes, wafer defect detection and re-inspection are core steps in ensuring chip yield and improving product reliability. Typically, defect detection equipment (such as optical inspection equipment or electron beam inspection equipment) performs a full scan of the wafer surface, identifying potential anomalies and generating an inspection report file (such as a standard Klarf file) containing the defect coordinates and some physical characteristic parameters. Subsequently, re-inspection equipment such as scanning electron microscopes (SEM) perform high-resolution imaging based on the defect coordinate list provided in the report, allowing engineers to observe the defect morphology in detail, analyze its properties and root causes, and then take corresponding process improvement measures.

[0003] However, the existing defect review process has significant technical limitations: 1. Currently, SEM machines typically receive and execute a list of defect coordinates transmitted by the defect management system. The generation of this list fails to effectively distinguish the potential severity of defects. A large amount of resources for scanning electron microscope review may be consumed in imaging "noise" defects (such as tiny non-critical particles, false defects, etc.) that have little impact on yield, while those few but potentially fatal defects that could directly lead to circuit failure are missed because they are not selected.

[0004] 2. When identifying defects, defect detection equipment not only records their coordinates but also captures a series of raw signal parameters reflecting the physical characteristics of the defects, such as defect size, threshold, and total energy / intensity. These parameters indirectly contain key information about the defect's material, height, and optical contrast, helping to predict defect types (e.g., distinguishing between particulate contamination, pattern indentation, or bridging anomalies). However, traditional sampling methods fail to fully integrate and utilize this rich physical feature information, relying solely on coordinates as the sampling basis. This results in a single dimension of judgment, making it impossible to accurately identify truly high-risk targets from a massive number of "suspected defects."

[0005] 3. Different regions on a chip have varying tolerances for defects. For example, defects located in core functional areas such as static random access memory (SRAM) cells are highly likely to cause memory cell failure, posing a direct threat to chip yield; while defects located in certain non-functional blank areas have negligible impact. Existing defect detection methods fail to differentiate the importance of defects in different regions, resulting in a failure to focus on defects in key areas during detection, leading to inaccurate defect identification.

[0006] In summary, existing defect sampling methods suffer from the aforementioned drawbacks, requiring engineers to process a large number of unverified and disorganized defect images during the SEM stage. This necessitates significant time-consuming manual browsing and screening, resulting in low efficiency. Furthermore, the risk of missing critical defects is high, hindering the rapid identification of process issues and the improvement of product yield. Therefore, there is an urgent need for a wafer defect sampling method that can quickly locate process problems and reduce the probability of missed defects. Summary of the Invention

[0007] This invention provides a wafer defect sampling method that can quickly locate process problems while reducing the probability of missed detections.

[0008] The present invention also provides a storage medium for implementing steps in a wafer defect sampling method.

[0009] One technical solution adopted in this invention is a wafer defect sampling method, comprising: acquiring a standard defect data file generated after a wafer defect inspection machine scans the wafer; parsing the standard defect data file to extract the coordinates and feature parameters of each defect to form a defect information set; mapping each defect to a specific region on the wafer based on the coordinates of each defect and wafer region information, wherein the specific region on the wafer includes a core region and a non-core region; sampling the defect information in the defect information set according to the type of feature parameters, prioritizing the sampling of defect information with higher parameter values; combining the information of all collected defects into a simplified defect list, and sending it to a SEM machine for scanning imaging and defect review.

[0010] The types of characteristic parameters include at least size, critical signal strength, total absorbed energy value, and energy concentration in the defect region. During the first sampling, at least 60 corresponding defect information items are collected for each defect characteristic parameter type in each round of sampling. For the defect information collected in each round, secondary sampling is performed according to the category of the defect region on the wafer, with the number of defect information items collected in the core region being greater than the number of defect information items collected in the non-core region.

[0011] The size value, critical signal strength value, total absorbed energy value, and energy concentration value of the defect area within the defect information set are divided into multiple sampling intervals in ascending order of value. During the first sampling, in each round of sampling, 60 defect information corresponding to each defect feature parameter type are collected from the two sampling intervals with the highest values. Before the second sampling, the number of defect information collected is no less than 300. During the second sampling, according to the set collection ratio of the core area and non-core area to which the defect belongs, the second sampling continues from the defect information collected in the first sampling until no less than 300 defect information are finally collected.

[0012] During the first sampling, if the number of defect information corresponding to each defect feature parameter type collected in the first round is less than 60, sampling will continue from other sampling intervals of each feature parameter type until 60 are collected.

[0013] If the total number of defect information after the first round of collection is less than 300, defect information that was not collected in the first round of collection will be collected from the defect information set according to the priority of the region to which the defect belongs, until 300 defect information are collected.

[0014] The sampling is performed according to the priority of the defect region, which includes dividing the wafer into a core functional region, a die region containing the core functional region, other die regions, and other regions on the wafer; the sampling priority of the core functional region, the die region containing the core functional region, other die regions, and other regions on the wafer decreases in that order.

[0015] The wafer region coordinates are defined in the wafer defect inspection machine configuration file. The wafer region coordinates are defined by importing the wafer design file or manually defining a specific region on the wafer when editing the wafer defect inspection machine configuration file. After the wafer defect inspection machine scans the wafer, it determines the specific location of the defect on the wafer based on the coordinates of the scanned wafer and outputs defect information with a location marker.

[0016] The regions defined on the wafer include the core functional region, the die region containing the core functional region, and other regions on the wafer. During secondary sampling, the sampling ratios of the core functional region, the die region containing the core functional region, other die regions, and other regions on the wafer are 45%-55%, 25%-35%, 10%-20%, and 5%-15%, respectively, and the sum of the sampling ratios of the four regions is 100%.

[0017] After performing SEM wafer re-inspection on the defect information included in the simplified defect list, an automatic defect classification system is used to classify the defect types, recording the defect types and defect feature parameters in different inspection levels into the defect database. When performing defect inspection on a new wafer, the acquired standard defect data file is parsed, the feature parameters of each defect are extracted, and based on the feature parameters of the defects, a defect type matching the feature parameters of the defects is searched in the defect database, and a defect type code is output. The output new wafer defect information set contains the coordinates, feature parameters, and defect type code of each defect.

[0018] Another technical solution of the present invention provides a computer storage medium storing a computer program thereon, characterized in that the computer program, when executed by a processor, implements the wafer defect sampling method described above.

[0019] The beneficial effects of this invention lie in its comparison of the physical characteristic signals of defects with the functional areas of the circuit layout. It performs an initial sampling based on multiple characteristic parameters such as the size, signal strength, and energy of the defect characteristic signals, and then performs a secondary sampling according to a "core functional area priority" weight. This ensures that the final defect list submitted for review is highly concentrated on those defects most likely to affect circuit function. This achieves a fundamental shift from blind general surveys to precise investigations. It greatly improves the detection rate of critical defects and the targeting of review, optimizing resource allocation from the root cause. Attached Figure Description

[0020] Figure 1 This is a flowchart of a wafer defect sampling method according to the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the present invention clearer, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Examples of these preferred embodiments are illustrated in the drawings. The embodiments of the present invention shown in and described with reference to the drawings are merely exemplary, and the present invention is not limited to these embodiments.

[0022] It should also be noted that, in order to avoid obscuring the invention with unnecessary details, only the structures and / or processing steps closely related to the solution according to the invention are shown in the accompanying drawings, while other details that are not closely related to the invention are omitted.

[0023] Example 1 Example 1 provides a wafer defect sampling method, including: acquiring a standard defect data file generated by a wafer defect inspection machine after scanning the wafer; parsing the standard defect data file to extract the coordinates and feature parameters of each defect to form a defect information set; mapping each defect to a specific region on the wafer based on the coordinates and wafer region information, wherein the specific region on the wafer includes a core region and a non-core region; sampling the defect information in the defect information set according to the type of feature parameters, prioritizing the sampling of defect information with higher parameter values; combining the information of all collected defects into a simplified defect list, and sending it to a SEM machine for scanning imaging and defect review.

[0024] Example 1 provides a wafer defect sampling method that processes a parsed standard defect data file. This file includes defect characteristic parameters, which characterize the type of defect and its impact on chip functionality. Example 1 partitions each defect characteristic parameter based on its type, selecting relevant parameters within the high-value range of each parameter. High values ​​indicate high-risk defects. The selected defect information (including coordinates and corresponding characteristic parameters) is packaged and sent to a SEM machine for scanning and verification to confirm its authenticity and importance. This invention, based on the analysis and selection of defect characteristic parameters, identifies important defects, improving the efficiency of SEM scanning and reducing the scanning of unimportant defects.

[0025] The following section provides a detailed explanation of a wafer defect sampling method in Embodiment 1. Figure 1 As shown, the wafer defect sampling method includes the following steps: Step 1: Obtain the standard defect data file and parse it.

[0026] Defect inspection equipment (such as optical inspection equipment or electron beam inspection equipment) performs a full scan of the wafer surface, identifies potential anomalies, and generates a standard defect data file (such as a standard Klarf file) containing the coordinates of the defects and some physical characteristic parameters.

[0027] The standard defect data file includes characteristic parameters of the wafer, which at least include size (the most intuitive physical indicator; the larger the size, the greater the impact on yield, such as causing short circuits or open circuits), critical signal strength (characterizing defect confidence; it is the threshold value for the machine to judge the strength of the defect signal; the larger the critical signal strength value, the higher the probability that the defect is real), total absorbed energy value (the scattering value or total absorbed energy value of the defect point, reflecting the significance of the defect; high-energy defects usually have higher scattering values ​​or total absorbed energy values), and energy concentration in the defect area (the concentration of signal energy in a specific pixel or small area at the location of the defect. It can be understood as the peak value of the "brightness" or "contrast" of that point in an image; this total energy value can characterize the signal strength of the defect).

[0028] Parse the standard defect data file to obtain the coordinate information and characteristic parameters of each defect (including at least the defect size, critical signal strength, total absorbed energy, and energy concentration in the defect area). Package the location information and characteristic parameters of all defects into a defect information set.

[0029] Step 2: Map the coordinates of each defect to a specific area of ​​the wafer and output the defect information with location markers.

[0030] Based on the coordinates of each defect and wafer region information, each defect is mapped to a specific region on the wafer, which includes a core region and a non-core region.

[0031] Specifically, the wafer region coordinates are defined in the wafer defect inspection machine configuration file. The wafer region coordinates are defined by importing the wafer design file or manually defining a specific region on the wafer when editing the wafer defect inspection machine configuration file (defined in the configuration file). Then, when the wafer defect inspection machine scans the wafer and detects a specific defect, it can identify the specific location of the defect on the wafer and output defect information with a location marker. This location marker can specifically identify the region where the defect is located in subsequent sampling rules.

[0032] The area information on the wafer includes the core area and non-core area of ​​the wafer. Specifically, the areas on the wafer are divided into core functional areas, die areas containing core functional areas, other die areas, and other areas on the wafer.

[0033] The core area includes core functional areas (such as SRAM storage areas or logic areas, which are more important than other areas on the wafer) and die areas containing the core functional areas (i.e., dies on the wafer; each die is a bare die. During the dicing process, the wafer is cut into multiple die areas, and each die becomes a separate chip after subsequent packaging processes. Each die has corresponding circuitry, and the core functional areas may exist on multiple die areas). Other die areas and other areas on the wafer are non-core areas.

[0034] Step 3: Sample information from the defect information set based on sampling rules.

[0035] Step 3-1: One sampling.

[0036] Step 3-1-1: First round of sampling.

[0037] The size values, critical signal intensity values, total absorbed energy values, and energy concentration values ​​of the defect location region within the defect information set are divided into multiple sampling intervals in ascending order of numerical value. For example, the defect size can be divided into four or more intervals: 0nm~20nm, 20nm~40nm, 40nm~60nm, and above 60nm.

[0038] In each round of sampling, 60 defect information entries corresponding to each defect feature parameter type are collected from the two sampling intervals with the highest values ​​(for example, when sampling based on defect size, the data can be collected in the interval above 60nm or in the interval between 40nm and 60nm). In each round of sampling, at least 60 corresponding defect information entries are collected for each defect feature parameter type.

[0039] To ensure that the number of samples collected in the first round meets the sampling requirements, if the number of defect information corresponding to each defect feature parameter type is less than 60 during the first round of sampling, sampling will continue from other sampling ranges for each feature parameter type until 60 defect information are collected. For example, when sampling based on defect size, if the number of samples collected in the range above 60nm or in the range of 40nm to 60nm is less than 60, sampling can be carried out from the range of 20nm to 40nm.

[0040] When performing secondary sampling, it is also necessary to ensure that the total number of defects in the first round of sampling is not less than 300. If the total number of defects after the first round of sampling is less than 300, defects that were not sampled in the first round of sampling should be supplemented from the defect information set according to the priority of the region to which the defect belongs.

[0041] The priority of the first round of supplementary acquisition can be set according to the regions on the wafer divided in step 2.

[0042] As an optional implementation, the sampling priority of the first round of supplementary acquisition is as follows: core functional area, die area containing core functional area, other die areas, and other areas on the wafer, with the sampling priority decreasing in that order.

[0043] Step 3-1-2: Multiple rounds of sampling.

[0044] Before each secondary sampling, it is necessary to ensure that the data from the first sampling is no less than 300. When the first sampling has collected 300 defect information, the secondary sampling will continue sampling based on the first sampling. Each round of secondary sampling will reduce the data in the database of the first sampling in each round. Therefore, it is necessary to replenish the database of the first sampling to ensure that the data in the database meets the basis of 300 defect information required for secondary sampling. Therefore, multiple rounds of sampling are to supplement defect information for secondary sampling.

[0045] Step 3-2: Secondary sampling.

[0046] In each sampling, the defect information collected in each round is sampled again according to the category of the defect region on the wafer. The number of defect information collected in the core region is greater than the number of defect information collected in the non-core region.

[0047] Specifically, based on the wafer region information obtained in step 2, the sampling ratios for the core functional region, the die region containing the core functional region, other die regions, and other regions on the wafer are 45%-55%, 25%-35%, 10%-20%, and 5%-15%, respectively, and the sum of the sampling ratios for the four regions is 100%.

[0048] After multiple rounds of secondary sampling, the secondary sampling ends when the number of defects collected reaches at least 300.

[0049] Step 4: Perform a follow-up SEM scan.

[0050] The information of all defects collected in step 3 is combined into a simplified defect list and sent to the SEM machine to scan and image the defects corresponding to all defect information and to conduct a defect review.

[0051] Step 5: After the SEM scan is completed, the defects are classified and the newly detected defect types are indicated.

[0052] After reviewing the defects included in the simplified defect list using SEM, the automatic defect classification system (ADC system, auto defect classification) is used to classify the defect types after review. The defect types and characteristic parameters of the defects in different detection levels (e.g., different metal layers are detected separately and defects in different layers are classified) are recorded in the defect database. When performing defect scanning on a wafer containing new defects, the wafer defect inspection machine generates a standard defect data file after scanning the new wafer, parses the standard defect data file, extracts the feature parameters of each defect, and searches for defect types that match the feature parameters in the defect database based on the defect feature parameters, and outputs the defect type code; the subsequent output defect information set contains the coordinates, feature parameters and type code of each defect. The defect type code can indicate the type of defect and suggest the risk level of the defect, which can be focused on during subsequent SEM review.

[0053] Example 2 Example 2 provides a computer storage medium on which a computer program is stored. When the computer program is executed by a processor, it implements a wafer defect sampling method as described in Example 1.

[0054] In the description of this invention, the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0055] Furthermore, it should be noted that in this specification, "comprising," "including," or any other variations thereof are intended to cover a 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. Without further limitations, 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 said element.

[0056] It should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A wafer defect sampling method, characterized in that, This includes acquiring a standard defect data file generated after a wafer defect inspection machine scans a wafer, parsing the standard defect data file, and extracting the coordinates of each defect and the characteristic parameters of the defect to form a defect information set. Based on the coordinates of each defect and wafer region information, each defect is mapped to a specific region on the wafer, which includes a core region and a non-core region. Based on the type of feature parameters, defect information within the defect information set is sampled separately, with priority given to collecting defect information with higher parameter values ​​during sampling. All the collected defect information is combined into a concise defect list and sent to an SEM machine for scanning and imaging, as well as for defect review.

2. The wafer defect sampling method according to claim 1, characterized in that, The types of characteristic parameters include at least size, critical signal strength, total absorbed energy, and energy concentration in the defect region; During the first sampling, at least 60 corresponding defect information entries were collected for each defect feature parameter type in each round of sampling; For each round of defect information collection, secondary sampling is performed according to the category of the defect region on the wafer. The number of defect information collected in the core region is greater than the number of defect information collected in the non-core region.

3. The wafer defect sampling method according to claim 2, characterized in that, The size value, critical signal strength value, total absorbed energy value, and energy concentration value of the area where the defect is located of all defects in the defect information set are divided into multiple sampling intervals in ascending order of numerical value. During the first sampling, in each round of sampling, 60 defect information corresponding to each defect feature parameter type were collected from the two sampling intervals with the highest values. Before conducting secondary sampling, the number of defect information collected shall not be less than 300; During the second sampling, the defect information collected in the first sampling is collected again according to the set collection ratio of the core area and non-core area to which the defect belongs, until no less than 300 defect information are finally collected.

4. The wafer defect sampling method according to claim 3, characterized in that, During the first sampling, if the number of defect information corresponding to each defect feature parameter type collected in the first round is less than 60, sampling will continue from other sampling intervals of each feature parameter type until 60 are collected.

5. The wafer defect sampling method according to claim 4, characterized in that, If the total number of defect information after the first round of collection is less than 300, defect information that was not collected in the first round of collection will be collected from the defect information set according to the priority of the region to which the defect belongs, until 300 defect information are collected.

6. The wafer defect sampling method according to claim 5, characterized in that, The sampling is carried out according to the priority of the area to which the defect belongs, including the wafer being divided into a core functional area, a bare die area containing the core functional area, other bare die areas, and other areas on the wafer; The sampling priority of the core functional area, the die area containing the core functional area, other die areas, and other areas on the wafer decreases in that order.

7. The wafer defect sampling method according to claim 1, characterized in that, Define the region coordinates of the wafer in the wafer defect inspection machine configuration file; The wafer's region coordinates are defined by importing the wafer design file or manually defining a specific region on the wafer when editing the wafer defect inspection machine configuration file. After scanning the wafer, the wafer defect testing machine determines the specific location of the defect on the wafer based on the coordinates of the wafer obtained from the scan, and outputs defect information with a location marker.

8. A wafer defect sampling method according to claim 7, characterized in that, The regions defined on the wafer include the core functional region, the die region containing the core functional region, and other regions on the wafer. During secondary sampling, the sampling ratios for the core functional area, the die area containing the core functional area, other die areas, and other areas on the wafer are 45%-55%, 25%-35%, 10%-20%, and 5%-15%, respectively, with the sum of the sampling ratios for the four areas being 100%.

9. A wafer defect sampling method according to claim 1, characterized in that, After performing SEM wafer re-inspection on the defect information included in the simplified defect list, the defect types are classified using an automatic defect classification system, and the defect types and characteristic parameters of the defects in different inspection levels are recorded in the defect database. When performing defect inspection on a new wafer, the acquired standard defect data file is parsed, the characteristic parameters of each defect are extracted, and based on the characteristic parameters of the defect, the defect type that matches the characteristic parameters of the defect is searched in the defect database, and the defect type code is output. The output set of new wafer defect information includes the coordinates, feature parameters, and defect type code for each defect.

10. A computer storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements a wafer defect sampling method as described in any one of claims 1-9.

Images