CN117197617A - A defect classification method and system for repeated defects - Google Patents

Abstract

The invention relates to the field of wafer inspection, and in particular to a defect classification method and system for repeated defects. The method includes: obtaining a position information set of at least one set of repeated defects of the wafer; obtaining a total sampling number and a sampling proportion relationship corresponding to the sampling positions of the wafer, so as to collect the total sampling from multiple pre-divided sampling areas on the wafer. The relationship between quantity and sampling ratio determines the number of regional samples in different sampling areas of the wafer; select corresponding sampling positions from multiple sampling areas according to the number of regional samples; take photos at the selected sampling positions to obtain A set of defect photos corresponding to at least one set of repeated defects is obtained; and the repeated defects are classified into defects through the set of defect photos. The present invention actually provides a standardized sampling process that can improve sampling automation, and the automated sampling method can perform low-correlation sampling in correlation areas, thereby ensuring the reliability of limited sample data.

Description

A defect classification method and system for repeated defects

Technical field

The invention relates to the field of wafer defect detection, and in particular to a defect classification method and system for repeated defects.

Background technique

In the production of semiconductor wafers, chemical vapor deposition, optical development, chemical mechanical polishing, etc. During this process, defects may occur on the wafer surface, and defects on the wafer will directly affect the working life and reliability.

Currently, the common classification process for wafer defects is:

Step 1: First use the defect scanning equipment to initially find the defects to be classified. For example, see CN115172199A which discloses a method and system for identifying wafer defects. This method uses the density distribution characteristics of clustered defects to extensively identify clustered defects (that is, cluster point defects) to guide subsequent accurate defect classification.

Step 2: Select a photography plan based on the scanning conditions of the defects to be classified in Step 1.

Specifically, when the number of identified defects is relatively small, the locations where the defects occur are photographed one by one to collect defect photos. For example, the invention patent application with publication number CN103344660A discloses an electron microscope analysis method for defect detection based on circuit patterns. This analysis method converts the defect location files obtained by preliminary scanning into defect files with characteristic circuit patterns. Then the electron microscope determines the defect locations by comparing the characteristic circuit patterns in the defect files, and then takes photos and samples of all defect locations one by one. . As another example, the invention patent application with publication number CN113013048A discloses a wafer defect detection method that uses array sampling rules for repeating units to conduct a relatively comprehensive sampling of wafer defects.

However, when the number of identified cluster point defects is too large, the collection pressure of defect images also increases significantly. At this time, usually only random points on the wafer can be photographed based on the current defect identification situation. That is, before production, staff set sampling inspection rules based on comprehensive decisions based on production process indicators (such as wafer design netlist) proposed by upstream customers, combined with factors such as defect distribution, quantity, and production management systems.

In addition, existing technologies also attempt to improve detection accuracy by splicing multiple photos. in,

The invention patent application with publication number CN115132599A discloses a defect detection method. In this method, multiple initial detection pictures are obtained by taking photos of the defects to be detected from multiple angles, and the final detection picture is obtained by splicing pictures to improve detection accuracy. However, this multi-picture splicing detection method undoubtedly requires a large amount of machine resources.

Step 3: Input all wafer scanning results (such as photo scanning results) into the automatic defect classification model (for example, ADC defect classification model) for automatic defect classification, or the staff can also perform manual classification.

In the current wafer defect detection process, the photography plan in step two is usually manually selected by engineers. However, the wafer production line usually runs 24 hours a day. Once a piece of equipment or a certain link fails, it will have a great impact on the delay of the entire production line. Therefore, factories have very high requirements for real-time defect detection efficiency. This method of manually selecting a photography plan has relatively serious problems in terms of time and machine resource usage. Especially when the process pattern of the wafer is relatively complex, and the types and distribution of defects are relatively wide, the time-consuming problem of step two will seriously affect the operation and maintenance of the entire production line.

For example, the invention patent application with publication number CN115015289A discloses an integrated circuit defect detection method. In this method, designers can detect different defects in integrated circuits according to actual conditions during the design stage of the chip design netlist (GDS). The functional importance of the chip is divided into regions, and each detection area is manually marked for subsequent photography and sampling. Subsequently, inspectors use automatic optical inspection equipment to compile inspection programs with different conditions for different inspection areas based on different marks. In other words, the above application proposes a method of manually marking functional detection partitions and differentially sampling different detection partitions. However, this method of manual partitioning and setting differentiated sampling conditions still has many problems when applied to actual wafer defect detection:

For example, on the one hand, on-site technical engineers and upstream designers are required to collaboratively participate in the process of selecting, marking, and analyzing sampling results, which obviously places extremely high demands on the labor and material costs in the wafer processing process. On the other hand, although this partition setting method based on differentiated sampling conditions reduces the detection time to a certain extent, the way of setting different detection conditions for different wafer manufacturing processes and different wafer detection areas makes the sampling results more reliable. Performance and accuracy are also easily affected by the professional capabilities of the staff.

Therefore, there is an urgent need for a sampling method that can improve defect sampling efficiency during the defect sampling stage while ensuring the reliability and accuracy of defect sampling.

Contents of the invention

The purpose of the present invention is to provide a defect classification method for repeated defects, which partially solves or alleviates the above-mentioned deficiencies in the prior art and can improve the efficiency of wafer defect classification.

In order to solve the above-mentioned technical problems, the present invention specifically adopts the following technical solutions:

A defect classification method for repeated defects, including steps:

S101 obtains a position information set of at least one group of repeated defects of the wafer, wherein the position information set includes: position information of at least one defect to be classified with the same or similar position;

S102 obtains the total sampling number and sampling proportion relationship corresponding to the sampling position of the wafer, so as to determine the location on the wafer according to the total sampling number and sampling proportion relationship from multiple pre-divided sampling areas on the wafer. The number of regional samples in different sampling areas of the circle; wherein the sampling position is used for photographing and sampling, and the wafer is divided into a first sampling area, a second sampling area and a third sampling area in the direction from its center to the edge. area;

S103 selects corresponding sampling positions from multiple sampling areas according to the number of regional samples; wherein, when the repeated defects include: a first repeated defect and a second repeated defect, accordingly, S103 includes the steps:

Selecting a first sampling position corresponding to the first repetitive defect from at least one sampling area;

Determine whether the first sampling position includes the second repetitive defect; if so, select at least one third sampling position containing the second repetitive defect in the sampling area where the first sampling position is located ;

When the distance between the third sampling position and the first sampling position is greater than the preset distance, the third sampling position is selected as the second sampling position;

S104: Take photos at the selected sampling positions to obtain a set of defect photos corresponding to at least one group of repeated defects;

S105 Classifies the repeated defects through the defect photo collection.

In some embodiments, before S102, the steps are also included:

S106 obtains the current processing technology of the wafer, and obtains historical data of wafer defects from the historical defect database through the current processing technology; wherein the historical defect database includes:

Defects formed on the wafer after at least one processing process, as well as the formation locations and causes of the corresponding defects; wherein the causes include one or more of the following: illumination factors, mechanical damage, edge effects, surface the remains;

S107 selects at least one annular area on the wafer; wherein, when the proportion of illumination factors among the causes of defects contained in the annular area falls within the first preset threshold range; and/or, when When the number of defects contained in the annular area that are not caused by illumination factors falls within the second preset threshold range, the annular area is determined as the second sampling area;

S108: Sequentially divide both sides of the second sampling area into a first sampling area and a third sampling area along the direction from the center to the edge of the wafer.

In some embodiments, before S106, the steps are also included:

S109 determines whether there is a processing technology matching the current processing technology in the area division database; wherein the area division database includes: at least one processing technology, and the division rule of the sampling area corresponding to the processing technology;

If so, select the division rule corresponding to the current processing technology, and use the corresponding division rule to divide the sampling area of the wafer;

If not, execute S106, or send a prompt signal to the user.

In some embodiments, the defect photo includes at least one complete functional unit or at least a group of related functional units.

In some embodiments, the width of the first sampling area, the radius of the second sampling area, and the width of the third sampling area are approximately 1:1:1.

In some embodiments, the sampling ratio relationship is: sampling locations are selected from the first sampling area, the second sampling area, and the third sampling area in a ratio of approximately 1:2:1.

In some embodiments, S105 includes the steps:

Input the set of defect photos into an ADC classification model, and output the defect category corresponding to at least one defect photo through the ADC classification model;

Wherein, when there is only one defect category with the largest number of defects, the defect category is used as the defect category of the repeated defects;

When the defect category with the largest number of defects is two or more, the defect category with the highest priority is selected as the corresponding defect category of the repeated defect, and/or the defect located in the second sampling area is selected. The defect category is used as the defect category corresponding to the repeated defect.

The present invention also provides a defect classification system for repeated defects, which includes:

The defect position acquisition module to be classified is configured to obtain a position information set of at least one group of repeated defects of the wafer, wherein the position information set includes: position information of at least one defect to be classified having the same or similar position;

A sampling data acquisition module configured to obtain the total sampling number and sampling proportion relationship corresponding to the sampling position of the wafer, so as to obtain the total sampling number from multiple pre-divided sampling areas on the wafer according to the total sampling number. The relationship between the sampling ratio and the sampling ratio determines the number of regional samples in different sampling areas of the wafer; wherein, the sampling position is used for photographing and sampling, and the wafer is divided into first sampling areas in sequence from its center to the edge. , the second sampling area and the third sampling area;

A sampling position acquisition module configured to select corresponding sampling positions from multiple sampling areas according to the regional sampling number; wherein, when the repeated defects include: a first repeated defect and a second repeated defect, Correspondingly, the sampling location acquisition module includes:

A first sampling unit configured to select a first sampling position corresponding to the first repetitive defect from at least one sampling area;

The second sampling unit is configured to determine whether the first sampling position includes the second repetitive defect; if so, select a sample containing the second repetitive defect in the sampling area where the first sampling position is located. At least one third sampling position for two repeated defects; when the distance between the third sampling position and the first sampling position is greater than the preset distance, the third sampling position is selected as the second sampling position;

A photographing module configured to take photographs at the selected sampling locations, thereby obtaining a set of defect photos corresponding to at least one group of repeated defects;

A classification module configured to perform defect classification on the repeated defects through the defect photo set.

In some embodiments, this includes:

The historical data acquisition module is configured to obtain the current processing technology of the wafer, and obtain historical data of wafer defects from the historical defect database through the current processing technology; wherein,

The historical defect database includes: defects formed after the wafer has been processed by at least one processing technology, as well as the formation location and cause of the corresponding defect; wherein the cause of formation includes one or more of the following: illumination factors, mechanical damage, edge effects, surface residues;

A first automatic partitioning module configured to select at least one circular ring area on the wafer;

Wherein, when the proportion of illumination factors among the causes of defects contained in the annular area falls within the first preset threshold range; and/or, when among the causes of formation of defects included in the annular area, When the number of defects that are not caused by illumination factors falls within the second preset threshold range, the annular area is determined as the second sampling area;

The second automatic partitioning module is configured to sequentially divide both sides of the second sampling area into a first sampling area and a third sampling area along the direction from the center to the edge of the wafer.

In some embodiments, the method further includes: a third automatic partition module configured to determine whether there is a processing technology matching the current processing technology in the area division database; wherein the area division database includes: at least one processing technology, and the division rule of the sampling area corresponding to the processing technology; if so, select the division rule corresponding to the current processing technology, and use the corresponding division rule to classify the sampling area of the wafer Carry out division; if not, output the division result to the sampling position acquisition module, or send a prompt signal to the user.

In some embodiments, the defect photo includes at least one complete functional unit or at least a group of related functional units.

Beneficial technical effects:

The present invention provides a method that can quickly sample and classify defects on wafers with numerous defect categories. Specifically, the present invention uses pre-grouped repetitive defects as sampling objects, and selects highly independent sample points for different repetitive defects in multiple associated areas (such as the first, second, and third sampling areas), so as to When the sample points are limited, the effectiveness of the sample point set should be improved as much as possible. In other words, the present invention can perform regional centralized sampling on the wafer scale and distributed sampling on the regional scale, thereby effectively reducing the number of necessary samples and improving the efficiency of the defect classification process.

Furthermore, the present invention can also select a limited combination of defect formation causes to quickly automate the setting of sampling partitions on the wafer, so as to reduce manual intervention in the sampling process to a certain extent and reduce the time necessary for defect classification. Moreover, through the selection of key defect causes such as lighting factors, mechanical damage, edge effects, and surface residues, wafers under typical wafer processing processes can be quickly partitioned, which is both accurate and versatile.

Description of the drawings

In order to more clearly explain the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to describe the embodiments or the prior art will be briefly introduced below. Throughout the drawings, similar elements or portions are generally identified by similar reference numerals. In the drawings, elements or parts are not necessarily drawn to actual scale. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without exerting any creative effort.

Figure 1 is a schematic flowchart of the steps of a wafer defect classification method in an exemplary embodiment of the present invention;

Figure 2 is a schematic diagram of defect distribution of a wafer in an exemplary embodiment;

Figure 3 is a schematic diagram of the sampling partition of the wafer in an exemplary embodiment;

Figure 4a is a schematic diagram of the selection of sampling locations in an exemplary embodiment;

Figure 4b is a schematic diagram of the selection principle of sampling positions in an exemplary embodiment;

Figure 5 is a schematic diagram of results of defect classification categories in an exemplary embodiment;

Figure 6 is a schematic diagram of the module structure of a defect classification method in an exemplary embodiment of the present invention.

Summary of reference signs:

01 is the wafer, 02 is the chip, 03 is the defect (or repeated defect), 03a is the first repeated defect,

03b is the second repetitive defect, 04 is the third sampling area, 05 is the second sampling area, 06 is the first sampling area, 07 is the sampling position, 07a is the first sampling position, 07b is the second sampling position, 07c is the Three sampling positions, 07d is the fourth sampling position.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

Herein, suffixes such as "module", "component" or "unit" used to represent elements are used only to facilitate the description of the present invention and have no specific meaning in themselves. Therefore, "module",

"Part" or "unit" may be used interchangeably.

In this article, the terms "upper", "lower", "inner", "outer", "front", "back", "one end", "other end", etc. indicate the orientation or positional relationship based on the orientation shown in the drawings. or positional relationships are only for the convenience of describing the present invention and simplifying the description, but do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance.

In this article, unless otherwise expressly stipulated and limited, the terms "installed", "provided with", "connected", etc. should be understood in a broad sense. For example, "connected" can be a fixed connection or a detachable connection, or Integrated connection; it can be mechanical connection, direct connection, indirect connection through an intermediary, or internal connection between two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood on a case-by-case basis.

As used herein, "and/or" includes any and all combinations of one or more of the associated listed items.

"Plural" in this article means two or more, that is, it includes two, three, four, five, etc.

It should be noted that, in this document, the terms "comprising", "comprises" or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article or device that includes a series of elements not only includes those elements, It also includes other elements not expressly listed or inherent in the process, method, article or apparatus. Without further limitation, an element defined by the statement "comprises a..." does not exclude the presence of additional identical elements in a process, method, article or apparatus that includes that element.

As used in this specification, the term "about" typically means +/-5% of the stated value,

More typically +/-4% of the stated value, more typically +/-3% of the stated value, more typically +/-2% of the stated value, even more typical of the stated value +/-1%, or even more typically +/-0.5% of the stated value.

In this specification, certain embodiments may be disclosed in a format that falls within a range. It should be understood that this "within a range" description is merely for convenience and brevity and should not be construed as an inflexible limitation on the scope of the disclosure. Accordingly, descriptions of ranges should be considered to have specifically disclosed all possible subranges and individual numerical values within such ranges. For example, range The description of Individual numbers, such as 1, 2, 3,

4, 5 and 6. The above rules apply regardless of the breadth of the scope.

In this article, "wafer" may also be called Wafer. Among them, wafer usually refers to the silicon wafer used to make silicon semiconductor circuits, and its original material is silicon.

In this article, "chip" may also be called bare crystal or die, that is, Die. Among them, Die is a small unpackaged integrated circuit body made of semiconductor materials. In other words, Die can refer to the die before the chip is packaged, which is a separate wafer area cut by laser from the wafer.

Generally, a Die can include a complete functional unit or a group of related functional units.

To facilitate subsequent production and assembly of integrated circuits. The functional unit may be a copy of a circuit, for example, it may be one or more tiny components such as a transistor, a capacitor, a resistor, etc. As shown in Figure 2, a wafer 01 can be cut into multiple Dies (chips 02) according to actual production requirements.

In this article, "defect points" usually refer to abnormal points on the wafer caused by processing technology problems, human errors, or other accidental factors (such as dust on the machine or wafer), which may directly affect Wafer operating life and reliability. Multiple closely spaced and coherent defect points in a wafer are generally considered to be clustered defect points. For example, multiple defect points connected without interruption at a point spacing of 150 microns can be regarded as a cluster of point defects (Cluster), also called a cluster (or wafer defect).

When these clusters are caused by process problems or human errors, they often have a certain degree of repeatability and will appear continuously on the next processed wafer, which has a greater impact on the yield of the wafer. Information about the size, geometry, and spatial location of clusters is therefore extremely valuable to process engineers seeking to identify potential production problems. Common clusters of point defects, such as mechanical damage, create regularly distributed wafer patterns.

In this article, "repeat defect" is also called: Repeater Defect or Repeater. in,

When cluster point defects (clusters) reoccur at the same or close locations on multiple dies of a wafer, or when cluster point defects reoccur at the same or close locations on multiple wafers, then Such cluster point defects can be determined as repeated defects. For example, repeated defects may be defects caused by corresponding process defects (such as dust, etc.) at the same location on the mask on the wafer. Moreover, when classifying defects, multiple cluster point defects with the same or similar positions can be regarded as the same group of repeated defects to be classified.

In this article, the mask is also called a photomask, photomask or photolithography mask, also known as Reticle or Mask. It is used as a pattern transfer tool or master in the microelectronics manufacturing process and carries the graphic design. and process technology information, considered the “negative” of the photolithography process.

Embodiment 1

As shown in Figures 1-5, the first aspect of the present invention provides a defect classification method for repeated defects.

Includes steps:

S101 obtains a position information set of at least a group of repeated defects of the wafer, where the position information set includes: position information of at least one defect to be classified having the same or similar position.

In some embodiments, the location information of multiple cluster point defects can be obtained through preliminary detection by a defect scanning device/defect detection device.

In some embodiments, the position information may be the coordinates of the clustered point defects on the wafer, or may also be the coordinates of the clustered point defects on the chip (Die).

In some embodiments, cluster point defects with the same or similar coordinates are grouped into the same group of repeated defects.

For example, in some embodiments, the defect detection equipment can generate a klarf file during the process of detecting wafer defects, and when parsing the klarf file, the equipment can perform Repeater calculations based on the location of the defect and assign the corresponding Repeater ID value. Among them, if the ID value is the same (except 0), it means that these Defects belong to the same Repeater group. Therefore, in this embodiment, Repeater information can be obtained from the defect database.

S102 obtains the total sampling number and sampling proportion relationship corresponding to the sampling position of the wafer, so as to determine the location on the wafer according to the total sampling number and sampling proportion relationship from multiple pre-divided sampling areas on the wafer. The number of regional samples in different sampling areas of the circle; wherein, the sampling position 07 is used for photographing and sampling, and the wafer is divided into a first sampling area 06 and a second sampling area 05 in the direction from its center to the edge. And the third sampling area 04, as shown in Figure 3.

Among them, the regional sampling number is: And n=1 represents the first sampling area,

n=2 represents the second sampling area, and n=3 represents the third sampling area.

In some embodiments, the total sampling number X and the sampling proportion relationship (Y1:Y2:Y3) can be preset values.

In some embodiments, the total sampling quantity and sampling proportion relationship can also be manually set or adjusted by users (such as process engineers) based on actual production needs.

Preferably, in some embodiments, for defect detection in typical processing processes, the ratio of the regional sampling numbers of the first sampling area, the second sampling area and the third sampling area is approximately: Y1:Y2:

Y3=1:2:1. In other words, the preferred sampling numbers for the first, second, and third sampling areas are approximately 0.25X,

0.5X, 0.25X.

S103 Select corresponding sampling positions from multiple sampling areas according to the area sampling number.

And as shown in Figure 4a-Figure 4b, when the repeated defects include: the first repeated defect 03a and the second repeated defect 03b, S103 includes the steps:

Select a first sampling position 07a corresponding to the first repetitive defect 03a from at least one sampling area;

Determine whether the first sampling position 07a includes the second repetitive defect 03b; if so,

Then, in the sampling area where the first sampling position 07a is located, the sample containing the second repeated defect is selected.

at least one third sampling position of trap 03b;

When the distance between the third sampling position and the first sampling position is greater than the preset distance,

Then the third sampling position is selected as the second sampling position 07b.

As shown in Figure 4b, in some embodiments, after the first sampling position 07a is determined in the first sampling area, it can be at any position in the first sampling area (such as inside the first sampling area, or in the first sampling area). (edge of the sampling area), select the second sampling position 07b, as long as the interval between the first and second sampling positions is greater than the preset distance L.

For another example, in some embodiments, when there are two or more groups of repeated defects on the chip where the sampling position is located, in order to quickly obtain the second sampling position, diagonal sampling, symmetrical sampling, etc. can also be used. Method to take points.

Specifically, in some embodiments, as shown in Figure 4b, the diagonal sampling method is to take the current first sampling position 07a (for example, it can be the center point coordinate of the sampling position) as point A, and take the wafer's Taking the center point as point B, we obtain a circle L1 whose center point is point B and whose circumference passes through point A. Taking point A as the vertex, set an inscribed polygon L2 (such as a rectangle, square, regular pentagon, etc.) in the circle L1, and check the positions of other vertices of the inscribed polygon. There will be a second repeated defect 03b. The vertex position or its adjacent area is selected as the second sampling position 07b.

Specifically, in some embodiments, the symmetrical sampling method is to take the current first sampling position 07a as point A, take the center point of the wafer as point B, draw a straight line passing through point A and point B, and Select point C on the straight line, where the distance between point C and point B is the same or similar to the distance between point A and point B. And when there is a second repetitive defect 03b at point C, point C or its adjacent area is selected as the second sampling position 07b.

It can be understood that the second sampling position selection rule in this embodiment can also be flexibly customized by the user based on actual needs, and the present invention is not limited here.

S104 Take photos at the selected sampling positions to obtain a set of defect photos corresponding to at least one group of repeated defects.

For example, in some embodiments, the sampling positions obtained in S103 can be transmitted to a scanning electron microscope (SEM), an optical microscope or other photographing equipment, and then the photographing equipment can take pictures and samples for each sampling position respectively.

S105 Classifies the repeated defects through the defect photo collection.

For example, in some embodiments, the defect photo set can be directly input into the ADC classification model for automatic defect classification, or the staff can also perform manual classification.

For wafer defect detection scenarios with many defect categories, this embodiment actually proposes a method of selecting low-correlation points in a correlation area (or the same area) to improve the accuracy of automated sample data collection. and reliability. Specifically, the present invention takes multiple repetitive defects obtained by pre-grouping as the selection object, and performs highly independent sample point selection for different repetitive defects in multiple associated areas (such as the first, second, and third sampling areas). , in order to improve the effectiveness of the sample point set as much as possible when the sample points are limited.

In some embodiments, before S102, the steps are also included:

S106 obtains the current processing technology of the wafer (for example, process type, process name, etc.),

And obtain historical data of wafer defects from the historical defect database through the current processing technology; wherein the historical defect database includes: defects formed on the wafer after at least one processing technology, and corresponding defects The formation location and formation reasons; wherein, the formation reasons include one or more of the following: light factors, mechanical damage, edge effects, surface residues;

S107 selects at least one annular area on the wafer; wherein, when the proportion of illumination factors among the formation causes contained in the annular area falls within the first preset threshold range; and/or, when the When the number of defects contained in the ring area that are not caused by illumination factors falls within the second preset threshold range, then the ring area is determined as the second sampling area;

S108: Sequentially divide both sides of the second sampling area into a first sampling area and a third sampling area along the direction from the center to the edge of the wafer.

In the embodiment of the present invention, limited typical factors such as illumination factors, mechanical damage, edge effects, and surface residues are comprehensively selected as reference conditions for accelerating partition setting, so as to facilitate rapid processing of wafers obtained by conventional wafer processing processes. Sampling partition settings. In other words, the partitioning method in this embodiment can further improve the automation of defect classification and reduce the necessity of manual intervention.

Preferably, in some embodiments, the edge effect includes one or more of the following: substrate edge effect, desorption edge effect, and thinning edge effect.

For example, in some embodiments, there may be problems such as lattice distortion and impurity accumulation at the edge of the substrate of the wafer, resulting in differences in the electrical performance of the wafer.

For example, in some embodiments, materials near the edge of the wafer may desorb due to changes in temperature, pressure, etc., thereby causing contamination of the wafer surface and affecting the electrical performance of the wafer.

For another example, in some embodiments, the thinning process of the wafer may cause a large error (or non-uniformity) in the thickness of the wafer material, thereby affecting the quality of the finished product of the wafer.

Preferably, when the number of repetitive defects initially detected is too large, or when the identification accuracy of defect types is required to be high, substrate edge effect, desorption edge effect, and thinning edge effect are selected as key edge effects in this embodiment. type to complete the automated sampling partition settings.

It can be understood that, unlike the existing method of manually selecting detection areas, the present invention proposes a standardized sampling location selection method that can reduce manual participation and improve sampling automation.

In some embodiments, before S106, the steps are also included:

S109 determines whether there is a processing technology matching the current processing technology in the area division database; wherein the area division database includes: at least one processing technology, and the division rule of the sampling area corresponding to the processing technology;

If so, select the division rule corresponding to the current processing technology, and use the corresponding division rule to divide the sampling area of the wafer;

If not, execute S106, or send a prompt signal to the user.

In some embodiments, the defect photo includes at least one complete functional unit or at least a group of related functional units.

In some embodiments, the width of the first sampling area, the radius of the second sampling area, and the width of the third sampling area are approximately 1:1:1.

In some embodiments, the width refers to the difference between the outer radius and the inner radius of the sampling area.

In some embodiments, the sampling ratio relationship is: sampling locations are selected from the first sampling area, the second sampling area, and the third sampling area in a ratio of approximately 1:2:1.

In some embodiments, S105 includes the steps:

Input the set of defect photos into an ADC classification model, and output the defect category corresponding to at least one defect photo through the ADC classification model;

Wherein, when there is only one defect category with the largest number of defects, the defect category is used as the defect category of the repeated defects;

When there are two or more defect categories with the largest number of defects, the defect type with the highest priority is selected as the corresponding defect type of the repeated defect, and/or the defect located in the second sampling area is selected. The defect type is used as the defect type corresponding to the repeated defect.

As shown in Figure 5, when the defect categories of the classified repeated defects are 1, 1, 2, and 3 respectively, the remaining unrecognized defect categories are also marked as 1.

When the defect categories of the classified repeated defects are 1, 1, 2, and 2 respectively. And at this time, the classification priority of defect category 2 is higher than that of defect category 1 (for example, the historical frequency of occurrence of defect category 2 is high), then the remaining unrecognized defect categories are also marked as 1.

In the embodiment of the present invention, a highly reliable defect photo set is collected by performing low-correlation sampling in the correlation area, and the above-mentioned rapid classification method can be further adopted for the above-described collection of defect photo sets to improve the quality of defects. Classification efficiency, so as to manage and maintain the wafer processing production line in a timely manner.

In some embodiments, bin represents a classification of defects (or, in other words, defect categories).

In some embodiments, the sampling position selected for the first sampling area may be within the first sampling area or may be at the edge of the first sampling area, such as the junction of the first and second sampling areas. Similarly, when sampling the second sampling area or the third sampling area, points can also be selected at their edges or boundaries.

Embodiment 2

As shown in Figure 6, the present invention also provides a defect classification system for repeated defects corresponding to the classification method in the above-mentioned Embodiment 1, which includes:

The defect position acquisition module 10 to be classified is configured to obtain the position information set of at least one group of repeated defects of the wafer, wherein the position information set includes: the position information of at least one defect to be classified having the same or similar position. ;

The sampling data acquisition module 20 is configured to acquire the total sampling number and sampling proportion relationship corresponding to the sampling position of the wafer, so as to obtain the total sampling number from multiple pre-divided sampling areas on the wafer according to the total sampling area. The relationship between quantity and sampling ratio determines the number of regional samples in different sampling areas of the wafer; wherein, the sampling position is used for photographing and sampling, and the wafer is divided into first sampling in the direction from its center to the edge. area, the second sampling area and the third sampling area;

The sampling position acquisition module 30 is configured to select corresponding sampling positions from multiple sampling areas according to the number of regional samples; wherein, when the repeated defects include: a first repeated defect and a second repeated defect , Correspondingly, the sampling position acquisition module includes: a first sampling unit 31 configured to select a first sampling position corresponding to the first repetitive defect from at least one sampling area; a second sampling unit 32 configured It is used to determine whether the first sampling position includes the second repetitive defect; if so, in the sampling area where the first sampling position is located, select at least one second repetitive defect that contains the second repetitive defect. Three sampling positions; when the distance between the third sampling position and the first sampling position is greater than the preset distance, the third sampling position is selected as the second sampling position;

The camera module 40 is configured to take pictures at the selected sampling positions,

thereby obtaining a set of defect photos corresponding to at least one group of repeated defects;

The classification module 50 is configured to perform defect classification on the repeated defects through the defect photo set.

In some embodiments, the system further includes:

The historical data acquisition module 60 is configured to obtain the current processing technology of the wafer, and obtain historical data of wafer defects from the historical defect database through the current processing technology; wherein, in the historical defect database, Including: defects formed after the wafer has been processed by at least one processing technology, as well as the formation location and causes of the corresponding defects; wherein the formation causes include one or more of the following: light factors, mechanical damage, edge Effects, surface residues;

The first automatic partitioning module 70 is configured to select at least one annular area on the wafer; wherein, when the proportion of illumination factors among the defect causes contained in the annular area belongs to the first preset within the threshold range; and/or when the number of defect causes that are not lighting factors among the defect causes contained in the annular area falls within the second preset threshold range, then the annular area is determined to be the second preset threshold range. sampling area;

a second automatic partitioning module 80 configured for use in a center-to-edge direction of the wafer,

In turn, both sides of the second sampling area are divided into a first sampling area and a third sampling area respectively.

In some embodiments, the system further includes:

The third automatic partitioning module is configured to determine whether there is a processing technology that matches the current processing technology in the area division database; wherein the area division database includes: at least one processing technology, and a processing technology that matches the current processing technology. The corresponding division rule of the sampling area; if yes, select the division rule corresponding to the current processing technology, and use the corresponding division rule to divide the sampling area of the wafer; if not, then divide The results are output to the sampling position acquisition module, or a prompt signal is sent to the user.

In some embodiments, the defect photo includes at least one complete functional unit or at least a group of related functional units.

In some embodiments, classification module 50 includes:

A first classification unit configured to input the defect photo set into an ADC classification model, and output the defect category corresponding to at least one defect photo through the ADC classification model;

Wherein, when there is only one defect category with the largest number of defects, the defect category is used as the defect category of the repeated defects;

The second classification unit is configured to select the defect type with the highest priority as the corresponding defect type of the repeated defect when the defect type with the largest number of defects is two or more;

And/or, a third classification unit configured to select the defect type of the defect located in the second sampling area as the corresponding defect type of the repeated defect.

It should be noted that, in this document, the terms "comprising", "comprises" or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article or device that includes a series of elements not only includes those elements, It also includes other elements not expressly listed or inherent in the process, method, article or apparatus. Without further limitation, an element defined by the statement "comprises a..." does not exclude the presence of additional identical elements in a process, method, article or apparatus that includes that element.

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 the necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is better. implementation. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product in essence or the part that contributes to the existing technology. The computer software product is stored in a storage medium (such as ROM/RAM, disk, CD), including several instructions to cause a computer terminal (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in various embodiments of the present invention.

The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above-mentioned specific implementations. The above-mentioned specific implementations are only illustrative and not restrictive. Those of ordinary skill in the art will Under the inspiration of the present invention, many forms can be made without departing from the spirit of the present invention and the scope protected by the claims, and these all fall within the protection of the present invention.

Claims (10)

1. A defect classification method for repeated defects, which is characterized by including the steps: S101 obtains a position information set of at least one group of repeated defects of the wafer, wherein the position information set includes: position information of at least one defect to be classified with the same or similar position; S102 obtains the total sampling number and sampling proportion relationship corresponding to the sampling position of the wafer, so as to determine the location on the wafer according to the total sampling number and sampling proportion relationship from multiple pre-divided sampling areas on the wafer. The number of regional samples in different sampling areas of the circle; wherein the sampling position is used for photographing and sampling, and the wafer is divided into a first sampling area, a second sampling area and a third sampling area in the direction from its center to the edge. area; S103 selects corresponding sampling positions from multiple sampling areas according to the number of regional samples; wherein, when the repeated defects include: a first repeated defect and a second repeated defect, accordingly, S103 includes the steps: Selecting a first sampling position corresponding to the first repetitive defect from at least one sampling area; Determine whether the first sampling position includes the second repetitive defect; if so, then In the sampling area where the first sampling position is located, at least one third sampling position containing the second repetitive defect is selected; When the distance between the third sampling position and the first sampling position is greater than the preset distance, the third sampling position is selected as the second sampling position; S104: Take photos at the selected sampling positions to obtain a set of defect photos corresponding to at least one group of repeated defects; S105 Classifies the repeated defects through the defect photo collection. 2. A defect classification method for repeated defects according to claim 1, characterized in that, before S102, it further includes the steps: S106 obtains the current processing technology of the wafer, and obtains historical data of wafer defects from a historical defect database through the current processing technology; wherein the historical defect database includes: the wafer has undergone at least one processing Defects formed after process treatment, as well as the formation locations and causes of corresponding defects; wherein the causes include one or more of the following: light factors, mechanical damage, edge effects, and surface residues; S107 selects at least one annular area on the wafer; wherein, when the proportion of illumination factors among the causes of defects contained in the annular area falls within the first preset threshold range; and/or, when When the number of defects contained in the annular area that are not caused by illumination factors falls within the second preset threshold range, the annular area is determined as the second sampling area; S108: Sequentially divide both sides of the second sampling area into a first sampling area and a third sampling area along the direction from the center to the edge of the wafer. 3. A defect classification method for repeated defects according to claim 2, characterized in that, before S106, it further includes the steps: S109 determines whether there is a processing technology matching the current processing technology in the area division database; wherein the area division database includes: at least one processing technology, and the division rule of the sampling area corresponding to the processing technology; If so, select the division rule corresponding to the current processing technology, and use the corresponding division rule to divide the sampling area of the wafer; If not, execute S106, or send a prompt signal to the user. 4. A defect classification method for repeated defects according to claim 1, characterized in that the defect photos include at least one complete functional unit or at least a group of related functional units. 5. A defect classification method for repeated defects according to claim 1, characterized in that the width of the first sampling area, the radius of the second sampling area, and the width of the third sampling area are approximately 1:1: 1; and/or, the sampling ratio relationship is: sampling positions are selected from the first sampling area, the second sampling area, and the third sampling area in a ratio of approximately 1:2:1. 6. A defect classification method for repeated defects according to claim 1, characterized in that S105 includes the steps: Input the set of defect photos into an ADC classification model, and output the defect category corresponding to at least one defect photo through the ADC classification model; Wherein, when there is only one defect category with the largest number of defects, the defect category is used as the defect category of the repeated defects; When the defect category with the largest number of defects is two or more, the defect category with the highest priority is selected as the corresponding defect category of the repeated defect, and/or the defect located in the second sampling area is selected. The defect category is used as the defect category corresponding to the repeated defect. 7. A defect classification system for repeated defects, characterized by including: The defect position acquisition module to be classified is configured to obtain a position information set of at least one group of repeated defects of the wafer, wherein the position information set includes: position information of at least one defect to be classified having the same or similar position; A sampling data acquisition module configured to obtain the total sampling number and sampling proportion relationship corresponding to the sampling position of the wafer, so as to obtain the total sampling number from multiple pre-divided sampling areas on the wafer according to the total sampling number. The relationship between the sampling ratio and the sampling ratio determines the number of regional samples in different sampling areas of the wafer; wherein, the sampling position is used for photographing and sampling, and the wafer is divided into first sampling areas in sequence from its center to the edge. , the second sampling area and the third sampling area; A sampling position acquisition module configured to select corresponding sampling positions from multiple sampling areas according to the regional sampling number; wherein, when the repeated defects include: a first repeated defect and a second repeated defect, Correspondingly, the sampling location acquisition module includes: A first sampling unit configured to select a first sampling position corresponding to the first repetitive defect from at least one sampling area; The second sampling unit is configured to determine whether the first sampling position includes the second repetitive defect; if so, select a sample containing the second repetitive defect in the sampling area where the first sampling position is located. At least one third sampling position for two repeated defects; when the distance between the third sampling position and the first sampling position is greater than the preset distance, the third sampling position is selected as the second sampling position; A photographing module configured to take photographs at the selected sampling locations, thereby obtaining a set of defect photos corresponding to at least one group of repeated defects; A classification module configured to perform defect classification on the repeated defects through the defect photo set. 8. A defect classification system for repeated defects according to claim 7, characterized in that it includes: The historical data acquisition module is configured to obtain the current processing technology of the wafer, and obtain historical data of wafer defects from the historical defect database through the current processing technology; wherein the historical defect database includes : Defects formed on the wafer after at least one processing process, as well as the formation location and cause of the corresponding defects; wherein the formation causes include one or more of the following: light factors, mechanical damage, edge effect , surface residue; The first automatic partitioning module is configured to select at least one annular area on the wafer; wherein, when the proportion of illumination factors among the causes of defects contained in the annular area belongs to the first predetermined When a threshold range is set; and/or when the number of defects contained in the annular area that are not caused by illumination factors belongs to the second preset threshold range, then the annular area is determined as second sampling area; The second automatic partitioning module is configured to sequentially divide both sides of the second sampling area into a first sampling area and a third sampling area along the direction from the center to the edge of the wafer. 9. A defect classification system for repeated defects according to claim 7, further comprising: The third automatic partitioning module is configured to determine whether there is a processing technology that matches the current processing technology in the area division database; wherein the area division database includes: at least one processing technology, and a processing technology that matches the current processing technology. The corresponding division rule of the sampling area; if yes, select the division rule corresponding to the current processing technology, and use the corresponding division rule to divide the sampling area of the wafer; if not, then divide The results are output to the sampling position acquisition module, or a prompt signal is sent to the user. 10. A defect classification system for repeated defects according to claim 7, characterized in that the defect photos include at least one complete functional unit or at least a group of related functional units.