WO2021142622A1

WO2021142622A1 - Method for determining cause of defect, and electronic device, storage medium, and system

Info

Publication number: WO2021142622A1
Application number: PCT/CN2020/072033
Authority: WO
Inventors: 王海金; 薛静
Original assignee: 京东方科技集团股份有限公司; 北京京东方光电科技有限公司
Priority date: 2020-01-14
Filing date: 2020-01-14
Publication date: 2021-07-22
Also published as: CN113597664A

Abstract

A method for determining the cause of a substrate defect. The method comprises: acquiring production process data of a substrate (S201), wherein the production process data comprises production history data of the substrate and parameter data of at least two production device parameters; acquiring, according to the type of the substrate, a parameter data reference range of the at least two production device parameters corresponding to the type of the substrate (S202); and on the basis of the acquired parameter data of the at least two production device parameters and the parameter data reference range of the at least two production device parameters, determining, from among the at least two production device parameters, a defective production device parameter that deviates from the parameter data reference range thereof (S203). By means of the method, the cause of a substrate defect can be quickly and conveniently evaluated.

Description

Method, electronic equipment, storage medium and system for determining the cause of bad

Technical field

The present invention relates to the field of panel production and processing, and more specifically, to a method, electronic equipment, computer-readable storage medium, and system for determining the cause of substrate failure.

Background technique

In the industrial production of chemicals and electronics, the qualification rate of each project is often referred to as the yield rate, which reflects the proportion of qualified products in the testing process of each project. The yield is directly related to the production cost. Whether the yield can be increased quickly in the shortest time determines whether the production cost can be recovered on time to a large extent. Yield rate, as a health indicator of factory products, has application value in all aspects of component manufacturing. A lower yield rate will lead to an increase in various costs. A high-level yield is a key indicator that reflects product reliability and realizes product revenue, and is particularly important in manufacturing enterprises in the component processing industry.

With the development of science and technology, the use of computers to process and analyze the data obtained in manufacturing, and the use of programs to help manual analysis, reasoning, judgment and decision-making are of great help to improving production efficiency and improving product yield.

Summary of the invention

According to a first aspect of the present disclosure, there is provided a method for determining the cause of the defective substrate (GLASS), including: obtaining substrate production process data, the production process data including substrate production history data and at least two production processes Parameter data of equipment parameters; obtaining, according to the type of the substrate, a parameter data reference range of the at least two production equipment parameters corresponding to the type of the substrate; and parameter data based on the obtained at least two production equipment parameters , And the parameter data reference range of the at least two production equipment parameters, determining the defective production equipment parameter that deviates from the parameter data reference range of the at least two production equipment parameters.

Optionally, the number of defective production equipment parameters that deviate from the parameter data reference range of the determined at least two production equipment parameters is at least two, and the method further includes: A substrate, using the substrate defective distribution pattern of each of the plurality of substrates to generate a defective distribution pattern corresponding to the production process data, wherein the defective distribution pattern shows the substrate coordinate position of the defective point; acquiring at least one A reference bad distribution pattern; and based on the bad distribution pattern and the at least one reference bad distribution pattern, determining the correlation ranking of the at least two production equipment parameters to be processed.

Optionally, the method further includes: acquiring the production process data of the panel samples of the plurality of substrates from a distributed storage device.

Optionally, the method further includes: displaying the defective production equipment parameters that deviate from the parameter data reference range of the at least two production equipment parameters that caused the occurrence of the defects, and/or displaying the sorted and determined defective production equipment Device parameters.

According to a second aspect of the present disclosure, an electronic device is provided. The electronic device includes a memory, a processor, and a computer program that is stored on the memory and can run on the processor, and is characterized in that, when the processor executes the program, the method for causing the panel failure as described above is implemented.

According to a third aspect of the present disclosure, there is provided a computer-readable storage medium, the computer-readable storage medium storing computer instructions, the computer instructions for causing the computer to perform the above-mentioned determination of the substrate failure Reason method.

According to a fourth aspect of the present disclosure, there is provided a system for determining the cause of substrate failure. The system includes a distributed storage device configured to store all production process data of all substrates within a preset time period; an electronic device as described above; and a display device configured to display the data to be displayed obtained from the electronic device Image data.

Description of the drawings

Through a more detailed description of the embodiments of the present disclosure in conjunction with the accompanying drawings, the above and other objectives, features, and advantages of the present disclosure will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present disclosure, and constitute a part of the specification, and are used to explain the present disclosure together with the embodiments of the present disclosure, and do not constitute a limitation to the present disclosure. In the drawings, the same reference numerals generally represent the same components or steps.

1A-1B show a flow chart of a method for determining the cause of substrate failure based on a yield analysis system;

FIG. 2 shows a schematic flowchart of a method for determining a cause of a substrate failure according to an embodiment of the present disclosure;

FIG. 3 shows a schematic flow chart of another method for determining the cause of a defective substrate according to an embodiment of the present disclosure;

4A-4B show schematic diagrams of displayed bad coordinate patterns and reference bad distribution patterns according to an embodiment of the present disclosure;

FIG. 5 shows a structural block diagram of an electronic device for determining the cause of a defective substrate according to an embodiment of the present disclosure; and

6A-6B show a structural block diagram of a system for determining the cause of a defective substrate according to an embodiment of the present disclosure.

Detailed ways

The technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by a person of ordinary skill in the art without creative work shall fall within the protection scope of the present disclosure.

Some products, for example, OLED (Organic Light-Emitting Diode) panels and LCD (Liquid Crystal Display, liquid crystal display) panels are highly integrated in the production process, from the initial glass substrate to the final panel cut from the substrate. There are many procedures, processes, and equipment, and there are many known and unknown defects in the production line. The defects of these production lines will be reflected in the final substrate to be cut into the panel. Moreover, defects are essentially caused by production equipment. Therefore, how to determine the cause of an abnormally high occurrence of defects in the final substrate is a problem that needs to be solved urgently.

The traditional determination of the cause of the failure is usually based on the Yield Management System (YMS).

The yield management system can obtain the data uploaded by the production equipment when the product passes through each production equipment, such as from the Manufacturing Execution System (MES) and the Fault Detection & Classification (Fault Detection & Classification, FDC) system. These data are collectively referred to as production data. The production data usually includes production process data and bad information data. The production process data includes: the relevant data of the production equipment parameters obtained from the FDC system, including the time from the start of the production of the product to the current time (the following uses the glass substrate (Glass) to generate the panel (Panel) as an example). Data measured by internal/external sensors in the production equipment (such as equipment operating time, temperature, humidity, pressure, number of vibrations, ratios of specific materials, etc.), that is, data related to production equipment parameters can refer to the operation of the production equipment/ State input data. In this data, the various physical/electrical state values of the production equipment are arranged in time series; the production history data obtained from the MES system, the production history data mainly includes historical data about the entire production, including historical data from the beginning The data of the executed process identifier (ID), material ID, equipment ID, equipment work specification ID, material ID, material ratio, etc. are arranged in a time series from the time of the start of the manufacture of the glass substrate to the current time. The bad information data is obtained from the MES system, and after all the processes of the glass substrate are completed and/or each process is completed separately, the inspection data for the entire glass substrate includes the substrate coordinates of many defective points Location information, and optional bad type information, etc. In addition, for the same batch of glass substrates, it is generally considered that the production is based on the same production process data (for example, the same production equipment and its parameters, the same process, the same material ratio, etc.), but each batch of glass substrates The bad information data of each substrate can be different.

The process of determining the cause of the defect based on the yield management system includes the following steps.

As shown in Figure 1A, when it is detected that an abnormally high incidence of defects (for example, multiple white spots and black spots) occurs on the final substrate to be cut into panels of a certain batch, the yield management system obtains the substrate in step S101 Corresponding production process data and bad information data. The yield management system has various defect analysis functions, such as the defect map analysis function. In step S102, based on the obtained defect information data, the substrate coordinate position of the defective point can be intuitively mapped to a single glass substrate, and The user interface of the yield management system graphically displays a substrate defect map (also including list information associated with each defect in the defect map) and/or a defect trend map on a single glass substrate. In step S103, the senior yield engineer checks the substrate defect map and/or the defect trend graph to understand where the defect is mainly concentrated, when, the trend of the bad coordinate, etc., and then can determine which specific production equipment is causing it based on experience High incidence of bad.

In addition, the above process can generally only locate equipment, but it may also be necessary to determine the specific production equipment parameters that cause the high incidence of defects. Therefore, as shown in FIG. 1B, after determining which specific production equipment is causing the high incidence of defects in step S103 based on experience, in step S104, each parameter data of the specific production equipment in the production process data is checked one by one to determine that it is the specific production equipment. The parameter data of which of the production equipment parameters is abnormal, which leads to a high incidence of defects.

Using the above process, after the final substrate has abnormally high incidence of defects, it takes a lot of time for the yield management system to run the defect map analysis function until the result to be displayed on the user interface is generated. In addition, if it is necessary to determine which production equipment parameter is abnormal, multi-level analysis is required. For example, after an engineer determines that equipment A is malfunctioning, he needs to further confirm each parameter of equipment A, which is complicated and inefficient. At the same time, the methods and types of data acquisition vary from person to person, and the data basis for decision-making is insufficient. Decision-making judgement depends entirely on the experience of personnel. The experience of bad cause analysis is not systematically solidified. The proportion of human factors is relatively large, and the requirements for yield analysis engineers Relatively high, if you are not an experienced engineer, you cannot correctly determine the cause of the failure.

Therefore, the embodiments of the present disclosure provide a new method, electronic device, computer-readable storage medium, and system for determining the cause of substrate failure. This method uses historical production process data to obtain the parameter data reference ranges of multiple production equipment parameters in advance, so that when defects on the substrate are high, the multiple production equipment corresponding to the defects can be quickly determined at this time. The parameter data of the parameter deviates from the parameter data reference range of the defective production equipment parameter. In addition, the method provided by the embodiment of the present disclosure can also obtain at least one reference defective distribution pattern in advance by using historical production process data and its corresponding defective information data, and make the substrate of each defective point on the at least one reference defective distribution pattern. The coordinate position is associated with its corresponding defective production equipment parameter, so that the substrate coordinate position of the defective point on the defective distribution pattern corresponding to the production process data in the actual production process at this time is correlated with at least one substrate that refers to the defective point on the defective distribution pattern The coordinate position is compared, based on the correlation between the substrate coordinate position of each defective point on at least one reference defective distribution pattern obtained in advance and its corresponding defective production equipment parameter, and the corresponding defective distribution pattern according to each defective production equipment parameter. The number of bad points is used to sort the correlation of each bad production equipment parameter with respect to the bad.

Using the method for determining the cause of the failure provided by the embodiments of the present disclosure, the user can more accurately and quickly determine the cause of the failure, take timely measures, and determine the correlation between the parameters of the poor production equipment and the failure. Sorting can be targeted to adjust the parameters of production equipment, while also reducing the experience and ability requirements of engineers.

Hereinafter, the embodiments of the present disclosure will be described in more detail with reference to FIGS. 2-5 and 6A-6B.

FIG. 2 is a schematic flowchart of a method 200 for determining a cause of a defective substrate according to an embodiment of the present disclosure. The method for determining the cause of the failure according to the embodiment of the present disclosure obtains the parameter data reference range of multiple production equipment parameters in advance by using historical production process data, so that it can quickly determine the time when the failure on the substrate is high. Among the parameter data of the multiple production equipment parameters corresponding to the bad, the bad production equipment parameter deviates from the reference range of the parameter data.

More specifically, as shown in FIG. 2, the method of determining the cause of the failure of the panel includes the following steps.

In step S201, the production process data of the substrate is acquired, where the production process data includes the production history data of the substrate and parameter data of at least two production equipment parameters.

Examples or explanations of the production history data and the parameter data of the production equipment parameters have been introduced in the foregoing, and the description will not be repeated here. Of course, those skilled in the art can also classify the production process data in other ways, and can also include other types of data.

In step S202, according to the type of the substrate, a parameter data reference range of at least two production equipment parameters corresponding to the type of the substrate is acquired.

Optionally, the parameter data reference ranges of the production equipment parameters corresponding to the different substrate types may be different.

Optionally, obtaining the parameter data reference range of at least two production equipment parameters corresponding to the type of the substrate includes: obtaining multiple sets of production process data of multiple substrate samples of the same type in the historical data, wherein the multiple substrate samples include the first A number of positive substrate samples and a second number of negative substrate samples; and based on the first number of positive substrate samples and their production process data, and the second number of negative substrate samples and their production process data, determine the at least two production The parameter data reference range of the device parameter.

Optionally, determining the parameter data reference range of the at least two production equipment parameters includes: training the production equipment using a first number of positive substrate samples and production process data, and a second number of negative substrate samples and production process data. Parameter model; and using the production equipment parameter model to generate at least two parameter data reference ranges for the production equipment parameters.

Optionally, a substrate includes multiple panels, and the positive and negative substrate samples can be distinguished according to the proportion of defective panels on each substrate. For example, the proportion of defective panels is less than 20%. Substrate samples, and the substrates with defective panels accounting for more than 20% are negative substrate samples. The ratio of positive and negative substrate samples may be the first quantity: the second quantity=7:3.

Optionally, the production equipment parameter model may be a convolutional neural network, linear regression, decision tree, random forest, support vector machine (SVM) or any other machine learning model.

Optionally, a test substrate sample can also be used to verify the reliability of the model. The test substrate sample can be imported into the trained production equipment parameter model, so as to test the trained production equipment parameter model, and the ratio of the test substrate sample to the training substrate sample can be 2:8. There are many ways to judge the quality of the model, such as using mean absolute error (mean_absolute_error), and cross validation error (cross_validation_error), and so on. The parameters and weights in the trained production equipment parameter model are further adjusted based on the judgment result of the measured substrate sample on the model quality (for example, the error back propagation algorithm used in the convolutional neural network model).

In step S203, based on the acquired parameter data of the at least two production equipment parameters and the parameter data reference range of the at least two production equipment parameters, determine the defective production equipment parameter that deviates from the parameter data reference range of the at least two production equipment parameters .

Optionally, the method may further include obtaining multiple sets of production process data of multiple substrate samples from the distributed storage device for training to obtain the production equipment parameter model. The related content of the distributed storage device will be described in detail later.

Therefore, through the above method for determining the cause of the substrate failure provided by the embodiments of the present disclosure, the parameter data reference range of the production equipment parameter is obtained in advance, so that the parameter data of the production equipment parameter in the actual production process data can be directly obtained. Compare with the reference range of the parameter data to quickly determine whether there is an abnormality in the production equipment parameter.

On the other hand, in the actual analysis process, it is also possible to sort the correlation between the parameters of each defective production equipment and the defects that occurred on the batch of substrates using the same production process data. FIG. 3 shows a flowchart of another method 300 for determining the cause of the defective substrate.

Obviously, only when the number of parameter data of bad production equipment parameters is at least two, the relevance ranking needs to be performed. The method 300 may include the following steps.

Steps S301-S303 are similar to S201-S203 in method 200, and will not be repeated here.

In step S304, for a plurality of substrates produced by using the production process data, a defective distribution pattern corresponding to the production process data is generated using the defective distribution pattern of each substrate in the plurality of substrates, wherein the defective distribution pattern shows defective points The coordinate position of the substrate.

Among them, the substrate defect distribution pattern of each substrate is obtained according to its corresponding defect information data. Bad information data can be obtained from the MES system, as described above.

Optionally, generating a defective distribution pattern corresponding to the production process data by using the defective distribution pattern of each substrate in the plurality of substrates includes: stacking the defective distribution patterns of the multiple substrates to obtain the defective distribution corresponding to the production process data pattern.

In step S305, at least one reference bad distribution pattern is acquired.

Optionally, acquiring at least one reference defective distribution pattern includes: acquiring a defective substrate distribution pattern of each defective substrate in a plurality of defective substrates in the historical data, wherein the defective substrate distribution pattern shows the substrate coordinate position of the defective point, and each The substrate coordinate positions of the defective points are associated with their corresponding defective production equipment parameters; and the defective substrate distribution patterns of the plurality of defective substrates in the historical data are stacked in at least one group to obtain at least one reference defective distribution pattern.

Optionally, the bad distribution pattern and at least one reference bad distribution pattern are displayed by a display device.

Optionally, the defective production equipment parameter associated with the defective point at the coordinate position of the substrate in the at least one reference defective distribution pattern can be determined by the user (for example, an experienced yield analysis engineer) according to the at least one reference defective distribution pattern. The coordinate position of the substrate is directly determined in advance, or the parameter data of the production equipment parameter in the production process data corresponding to the defective point at the coordinate position of the substrate in the at least one reference defective distribution pattern can be compared with the parameter data reference range to obtain The defective production equipment parameter that exceeds the reference range of the parameter data is taken as the defective production equipment parameter associated with the defective point at the coordinate position of the substrate in the at least one reference defective distribution pattern.

In step S306, based on the bad distribution pattern and at least one reference bad distribution pattern, determine the correlation ranking of each bad production equipment parameter and the bad.

Optionally, based on the bad distribution pattern and at least one reference bad distribution pattern, determining the correlation ranking of each bad production equipment parameter and the bad includes: for each bad point in the bad distribution pattern, determining the substrate coordinate position of the bad point, And obtain the defective production equipment parameter associated with the defective point at the coordinate position of the substrate in at least one reference defective distribution pattern; for each determined defective production equipment parameter, determine the number of corresponding defective points; and The number of defective points corresponding to each of the determined defective production equipment parameters is sorted by correlation for the determined defective production equipment parameters.

Optionally, based on the bad distribution pattern and at least one reference bad distribution pattern, determining the correlation ranking of each bad production equipment parameter and the bad includes: for each bad point in the bad distribution pattern, determining the substrate coordinate position of the bad point, And obtain the defective production equipment parameter associated with the defective point at the coordinate position of the substrate in at least one reference defective distribution pattern; for each determined defective production equipment parameter, determine the number of panels distributed by the corresponding defective point ; And according to the number of panels distributed by the respective defective points corresponding to the determined defective production equipment parameters, the determined defective production equipment parameters are sorted in correlation.

In addition, in order to better improve the user interaction experience so that the user can adjust the parameter data of the bad production equipment parameter, the method 300 may further include step S307: displaying the deviation of the parameter data of the at least two production equipment parameters that caused the failure. The defective production equipment parameters in the reference range, and/or display the sorted and determined defective production equipment parameters.

Optionally, the above-mentioned content is displayed to the user through a display device.

In addition, it should be understood that since the historical data includes a large amount of production process data and bad information data of defective substrates over a long period of time, it can be considered that the coordinate positions of the defective points in the defective distribution pattern in the actual production process are referred to the defective distribution. There are also bad spots on the pattern here.

It is worth noting that, in an alternative embodiment, for example, the correlation between the production equipment parameter that causes the substrate defect and the defect can be directly obtained, so step S302 to step S303 can be omitted. That is to say, after obtaining the production process data of the substrates on which defects are high, proceed directly to step S304, that is, use the substrate of each substrate in the plurality of substrates (for example, a batch of substrates) corresponding to the production process data. The defective distribution pattern generates a defective distribution pattern corresponding to the production process data, wherein the defective distribution pattern shows the substrate coordinate position of the defective point, and then steps S305 to S307 are similarly performed.

The following illustrates an example process of a method for determining a cause of a substrate failure according to an embodiment of the present disclosure. 4A-4B show schematic diagrams of displayed bad coordinate patterns and reference bad distribution patterns according to an embodiment of the present disclosure.

When it is found that there is a high incidence of defects on a batch of substrates that have gone through all the processes but have not been cut into panels, the production process data of this batch of substrates can be obtained from the MES system and the FDC system. The production process data includes the production history of the substrates Data (for example, process 1, process 2, material type, material ratio) and parameter data of at least two production equipment parameters (for example, the operating parameters of each production equipment, including temperature, humidity, pressure, working time, etc.). At the same time, the bad information data of each substrate (such as the bad coordinate position of the final substrate 1 (Glass 1), the bad coordinate position of the final substrate 2 (Glass 2), etc.) can also be obtained from the MES system.

Compare the parameter data of at least two production equipment parameters with the known parameter data reference ranges of at least two production equipment parameters, and determine the defective production equipment in the reference data of the at least two production equipment parameters that deviates from its parameter data reference range Parameters, for example, determine that, for example, equipment 1-temperature, equipment 1-pressure, equipment 2-temperature, equipment 2-humidity, equipment 3-pressure, and equipment 4-humidity exceed their reference ranges. Wherein, the parameter data reference range of the known at least two production equipment parameters is obtained by training in advance based on the production process data of multiple historical substrate samples, wherein the substrate samples include a first number of positive substrate samples and a second number of substrate samples. Negative substrate sample.

Then, when a yield analysis engineer who is not experienced enough needs to analyze the correlation between the parameter data of the current bad production equipment parameters and the occurrence of the bad, he can continue or alternatively perform the following operations. For this batch of substrates produced with the same production process data, the substrate coordinate position of the defective point on the substrate defective distribution pattern of the substrate 1 is (20,20-panel 1) (15,30-panel 1), and the position of the substrate 2 The substrate coordinate position of the defective point on the substrate defective distribution pattern is (20,40-panel 1) (55,60-panel 2), and the substrate coordinate position of the defective point on the substrate defective distribution pattern of the substrate 3 is (100,100-panel 5) (120, 150-Panel 6), etc., stack the defective distribution pattern of each substrate in the multiple substrates of this batch to generate the defective distribution pattern (MAP R) corresponding to the production process data. The defective distribution pattern (MAP R) shows the coordinate positions of all defective points on the substrate, that is, the size of the defective distribution pattern corresponds to the size of the substrate, and the defective points are located at (20,20)(15,30)(20,40). )(55,60)(100,100)(120,150) and so on at the coordinate position of the substrate, as shown in Figure 4A. Of course, these specific coordinate positions of the substrate are only examples, and other positions of the substrate may also have defects.

For the defective point at the substrate coordinate position (20, 20) in the defective distribution pattern, obtain the known defective point at the substrate coordinate position (20, 20) in the at least one reference defective distribution pattern The associated bad production equipment parameters, and it is known that the temperature of equipment 1 and the humidity of equipment 2 caused the failure to occur.

For the defective point at the substrate coordinate position (15, 30) in the defective distribution pattern, obtain the known defective point at the substrate coordinate position (15, 30) in the at least one reference defective distribution pattern The associated bad production equipment parameters, and it is known that the humidity of equipment 1 and the pressure of equipment 2 caused the failure to occur.

For the defective point at the substrate coordinate position (20, 40) in the defective distribution pattern, obtain the known defective point at the substrate coordinate position (20, 40) in the at least one reference defective distribution pattern The associated bad production equipment parameters, and it is known that the temperature of equipment 1 and the pressure of equipment 2 caused the failure to occur.

For the defective point at the substrate coordinate position (55, 60) in the defective distribution pattern, obtain the known defective point at the substrate coordinate position (55, 60) in the at least one reference defective distribution pattern Related bad production equipment parameters, and it is known that the temperature of equipment 1 and the humidity of equipment 4 caused the failure to occur.

For the defective point at the substrate coordinate position (100, 100) in the defective distribution pattern, obtain the known defective point associated with the defective point at the substrate coordinate position (100, 100) in the at least one reference defective distribution pattern. Production equipment parameters, knowing that the temperature of equipment 2 and the pressure of equipment 3 caused the failure to occur.

For the defective point at the substrate coordinate position (120, 150) in the defective distribution pattern, obtain the known defective point associated with the defective point at the substrate coordinate position (120, 150) in the at least one reference defective distribution pattern. Production equipment parameters, knowing that the pressure of equipment 3 and the humidity of equipment 4 caused the failure.

By performing the above operations, the number of defective points corresponding to each defective production equipment parameter can be obtained, that is, the number of occurrences of each defective production equipment parameter on the defective distribution pattern, so as to sort the determined defective production equipment parameters. More specifically, for several exemplary substrate coordinate positions in the above-mentioned bad distribution pattern, device 1-temperature affects three bad points (20, 20), (20, 40) and (55, 60), device 1- Humidity affects (20,40) a bad point, equipment 2-pressure affects (15,30) and (20,40) two bad points, equipment 2-temperature affects (20,20) a point, equipment 2- Temperature affects one point (100,100), equipment 3-pressure affects (100,100) and (120,150) two bad points, equipment 4-humidity affects (55,60) and (120,150) two bad points. Therefore, it can be determined that the correlation order of the various bad production equipment parameters that lead to the occurrence of the substrate failure is: equipment 1-temperature; equipment 2-pressure, equipment 4-humidity (parallel); equipment 2-humidity, equipment 3-pressure, Equipment 1-Humidity (tie).

The above sorting is based on the number of bad points corresponding to each bad production equipment parameter. Of course, as mentioned above, it can also be sorted according to the number of panels distributed by bad points corresponding to each bad production equipment parameter. This will not be described in detail here.

In addition, the process of obtaining at least one reference bad distribution pattern is: obtaining the bad distribution pattern of each bad substrate in the multiple bad substrates (Glass 1,..., Glass n-1, Glass n) on which there are bad substrates in the historical data , Wherein each substrate defective distribution pattern shows the substrate coordinate position of one or more defective points on it, and the substrate coordinate position of each defective point is associated with its corresponding defective production equipment parameter; and the historical data The substrate defective distribution patterns of multiple defective substrates are stacked according to at least one group. Taking one group as an example, a reference defective distribution pattern (MAP REF) is obtained. The reference defective distribution pattern (MAP REF) includes all the substrates on the multiple substrates. The substrate coordinate position of the defective point.

The defective production equipment parameters associated with the substrate coordinate position of each defective point are pre-calculated and stored, so that they are known, and the acquisition process is: display a sample reference defective distribution pattern on the display device, as shown in Figure 4B As shown, the coordinate position of each substrate corresponding to each defective point of the defective distribution pattern as shown in FIG. 4A is determined through comparison, and then the user according to the coordinate position of each substrate in the reference defective distribution pattern, Compare the parameter data of the production equipment parameters in the production process data corresponding to the defective point at each substrate coordinate position in the reference defective distribution pattern with the parameter data reference range, and obtain the correlation with the substrate coordinate position of each defective point Bad production equipment parameters. Therefore, the defective production equipment parameters associated with the substrate coordinate position of each defective point are pre-calculated and stored. When the actual defective distribution pattern is obtained, the operation as described above is performed on the defective distribution pattern. Known, defective production equipment parameters associated with the substrate coordinate position of each defective point are used as the basis of the analysis.

By adopting the above-mentioned method, by obtaining the reference defective distribution pattern and the defective production equipment parameters associated with the coordinate position of the substrate of each defective point in advance, the actual defective distribution pattern is coordinated with the reference defective distribution pattern to obtain the actual The correlation of the defective production equipment parameters associated with each defective point in the defective distribution pattern is determined by determining the number of defective points caused by each defective production equipment parameter. Therefore, it is convenient for the less experienced yield analysis engineers to quickly check Sort the correlation between the equipment parameters and the defects that caused the substrate defects.

At the same time, in the embodiments of the present disclosure, the defective distribution pattern and the substrate coordinate position of the defective point in the reference defective distribution pattern are described in a graphical manner, but the text form may also be adopted. For example, the substrate coordinate positions of all defective points on multiple defective substrates on which defective substrates exist in the historical data are stored in text form. When the actual substrate coordinate positions of the defective points in the actual production process are obtained, each actual The substrate coordinate position and the stored substrate coordinate position are compared in text to determine which of the stored substrate coordinate positions the actual substrate coordinate position corresponds to, and the corresponding defective production equipment parameters are obtained.

According to another aspect of the present disclosure, an electronic device 500 is also provided.

FIG. 5 shows a structural block diagram of an electronic device 500 according to an embodiment of the present disclosure. The electronic device 500 includes a memory 501 and a processor 502. A computer program is stored on the memory, and when the computer program runs on the processor, the method for determining the cause of the substrate failure as described with reference to FIGS. 2 to 3 is implemented.

According to another aspect of the present disclosure, there is also provided a system 600 for determining the cause of substrate failure.

6A-6B show a structural block diagram of a system 600 according to an embodiment of the present disclosure. As shown in FIG. 6A, the system includes a distributed storage device 610, an electronic device 620 (for example, the electronic device 500 described with reference to FIG. 5), and a display device 630.

The distributed storage device 610 is configured to store all production process data and bad information data of all substrates within a preset time period. For example, it is possible to store all production process data and bad information data corresponding to all substrates in the process of producing all substrates within two years.

Among them, the distributed storage device stores relatively complete data (such as a database), and the distributed storage device includes multiple hardware memories, and different hardware memories are distributed in different physical locations (such as in different factories, or in different Different production lines), and realize the transfer of information between each other through the network, so that the data is distributed, but logically constitute a database based on big data technology.

Referring to Figure 6B, the raw data of a large number of different factory equipment are stored in the corresponding manufacturing system, such as the relational system of yield management system (YMS), error detection and classification (FDC) system, manufacturing execution system (MES), etc. Database (such as Oracle, Mysql, etc.), and these raw data can be extracted from the original table by data extraction tools (such as Sqoop, kettle, etc.) to be transmitted to distributed storage devices (such as Hadoop Distributed File System, HDFS, Hadoop Distributed File System, HDFS) ), in order to reduce the load on the factory equipment and manufacturing system, and facilitate the subsequent data reading of the analysis equipment.

The data in the distributed storage device can be stored in Hive tool or Hbase database format. For example, according to the Hive tool, the above raw data is first stored in the data lake; later, in order to reduce the learning cost of data cognition and the unity of business realization, you can continue to perform data in the Hive tool according to the application theme and scenario of the data. Preprocessing such as cleaning and data conversion to obtain data warehouses with different themes (for example, production history data theme, bad detection data theme, bad point measurement data theme, and production equipment parameter data theme), and different scenarios (such as sudden Data marts for bad sex scenes and correlation analysis scenes. It should be noted that the data topics and scenarios are not limited to the above examples. New data topics and scenarios can be added or reconstructed according to business needs. For example, the subject of bad detection data and the subject of bad point measurement data can be merged into the production history data. Topic. The above data mart can be connected to display devices, electronic devices, etc. through different API interfaces to realize data interaction with these devices.

Among them, due to multiple factory equipment in multiple factories, the data volume of the above raw data is very large. For example, the raw data generated by all factory equipment every day may be several hundred G, and the data generated every hour may also be tens of G.

There are two main solutions for the storage and calculation of massive structured data: RDBMS relational database management (Relational Database Management System, RDBMS) grid computing solution; distributed file management system (Distributed File System, DFS) big data solution .

Among them, the grid computing of RDBMS divides the problem that requires very huge computing power into many small parts, and then distributes these parts to many computers for separate processing, and finally combines these calculation results. For example, as a specific example, Oracle RAC (Real Application Cluster) is the core technology of grid computing supported by the Oracle database, in which all servers can directly access all data in the database. However, the RDBMS grid computing application system cannot meet user requirements when the amount of data is large. For example, due to the limited expansion space of the hardware, when the data is increased to a large enough order of magnitude, the input/output bottleneck of the hard disk will cause The efficiency of processing data is very low.

Big data technology based on distributed file management allows the use of multiple inexpensive hardware devices to build large clusters to process massive amounts of data. For example, the Hive tool is a Hadoop-based data warehouse tool that can be used for data extraction, transformation and loading (ETL). The Hive tool defines a simple SQL-like query language, and it also allows custom MapReduce mappers and reducers to be used by default tools. Complex analysis work. The Hive tool does not have a special data storage format, nor does it create an index for the data. Users can freely organize the tables in it and process the data in the database. It can be seen that the parallel processing of distributed file management can meet the storage and processing requirements of massive data. Users can process simple data through SQL queries, and use custom functions for complex processing. Therefore, when analyzing the massive data of the factory when all the substrates are produced, it is necessary to extract the data of the factory database into the distributed file system, which will not damage the original data on the one hand, and improve the efficiency of data analysis on the other hand.

The electronic device 620 may include one or more processors, and the one or more processors are configured to perform operations for determining the correlation. The electronic device 620 includes a processor (such as a CPU) with data processing capabilities, and may also have a memory (such as a hard disk) storing required programs. The processor and the memory are connected through I/O to realize information interaction, so that the processor The required operation can be performed according to the program stored in the memory to realize the operation of determining the correlation.

The electronic device 620 may be the electronic device 500 described with reference to FIG. 5.

The display device 630 has a display function for displaying the image data to be displayed obtained from the electronic device. According to an embodiment of the present disclosure, the display data may include a bad distribution pattern and at least one reference bad distribution pattern, a bad production equipment parameter that deviates from its parameter data reference range among a plurality of production equipment parameters that causes a bad occurrence, and/or a display experience Sort the determined parameters of the bad production equipment.

The display device 630 may include one or more displays, including one or more terminals with display functions, so that the electronics can send the calculated display data to the display device, and the display device will then display it.

In some embodiments, the display device can also be used to display an "interactive interface", which can include a sub-interface that displays the determined result (such as poor production equipment parameters, correlation), and is used to control the system to perform the required work ( Such as task setting) sub-interface, as well as the sub-interface for controlling each production equipment (such as modifying its production equipment parameters).

In other words, through the "interaction interface" of the display device, the user can fully interact with the system that determines the cause of the substrate failure (control and receive results).

In the embodiments of the present disclosure, the distributed storage device can efficiently realize the collection and preliminary processing of the raw data of multiple production equipment through big data, and the electronic device can easily obtain the required data from the distributed storage device. The defective production equipment parameters beyond the reference range of the parameter data of the production equipment parameters can be calculated by calculation, and the correlation between the defective production equipment parameters and the bad can be further obtained and displayed by the display device. Therefore, the embodiment of the present disclosure can automatically determine the cause of the substrate defect, so as to locate the cause of the defect, adjust the production process, and so on.

According to another aspect of the present disclosure, a computer-readable storage medium is also provided. The computer-readable storage medium stores computer instructions for causing the computer to execute the method for determining the cause of the defective substrate as described with reference to FIGS. 2 to 3.

Although the subject matter has been described in detail for various specific example embodiments of the subject matter, each example is provided by way of explanation and not limitation of the present disclosure. Those skilled in the art can easily make changes, changes and equivalents of such embodiments after gaining an understanding of the above content. Therefore, the present invention does not exclude the inclusion of such modifications, changes and/or additions to the subject matter that will be obvious to those of ordinary skill in the art. For example, features illustrated or described as part of one embodiment can be used with another embodiment to yield a still further embodiment. Therefore, it is intended that the present disclosure covers such alterations, changes and equivalents.

This application uses specific words to describe the embodiments of this application. For example, "one embodiment", "an embodiment", and/or "some embodiments" mean a certain feature, structure, or characteristic related to at least one embodiment of the present application. Therefore, it should be emphasized and noted that “one embodiment” or “one embodiment” or “an alternative embodiment” mentioned twice or more in different positions in this specification does not necessarily refer to the same embodiment. . In addition, some features, structures, or characteristics in one or more embodiments of the present application can be appropriately combined.

Those skilled in the art can understand that various aspects of this application can be illustrated and described through a number of patentable categories or situations, including any new and useful process, machine, product or combination of substances, or any combination of them. New and useful improvements. Correspondingly, various aspects of the present application can be completely executed by hardware, can be completely executed by software (including firmware, resident software, microcode, etc.), or can be executed by a combination of hardware and software. The above hardware or software can all be referred to as "data block", "module", "engine", "unit", "component" or "system". In addition, various aspects of the present application may be embodied as a computer product located in one or more computer-readable media, and the product includes computer-readable program codes.

Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which this disclosure belongs. It should also be understood that terms such as those defined in ordinary dictionaries should be interpreted as having meanings consistent with their meanings in the context of related technologies, and should not be interpreted in idealized or extremely formalized meanings, unless explicitly stated here. So defined.

The above is an explanation of the present disclosure, and should not be considered as a limitation to it. Although several exemplary embodiments of the present disclosure have been described, those skilled in the art will readily understand that many modifications can be made to the exemplary embodiments without departing from the novel teachings and advantages of the present disclosure. Therefore, all these modifications are intended to be included in the scope of the present disclosure defined by the claims. It should be understood that the above is an explanation of the present disclosure and should not be considered as being limited to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included in the scope of the appended claims. The present disclosure is defined by the claims and their equivalents.

Claims

A method to determine the cause of the substrate failure, including:

Acquiring production process data of the substrate, where the production process data includes production history data of the substrate and parameter data of at least two production equipment parameters;

Acquiring, according to the type of the substrate, the parameter data reference range of the at least two production equipment parameters corresponding to the type of the substrate; and

Based on the acquired parameter data of the at least two production equipment parameters and the parameter data reference range of the at least two production equipment parameters, determine the defective production equipment parameter that deviates from the parameter data reference range of the at least two production equipment parameters .
The method according to claim 1, wherein the number of defective production equipment parameters that deviate from the parameter data reference range of the determined at least two production equipment parameters is at least two,

The method also includes:

For a plurality of substrates produced using the production process data, use the substrate defective distribution pattern of each of the plurality of substrates to generate a defective distribution pattern corresponding to the production process data, wherein the defective distribution pattern shows The coordinate position of the substrate of the defective point;

Obtain at least one reference bad distribution pattern; and

Based on the bad distribution pattern and the at least one reference bad distribution pattern, a correlation ranking of each bad production equipment parameter and the bad is determined.
The method according to claim 1, wherein the obtaining, according to the type of the substrate, the parameter data reference range of the at least two production equipment parameters corresponding to the type of the substrate further comprises:

Acquiring multiple sets of production process data of multiple substrate samples of the same type in the historical data, wherein the multiple substrate samples include a first number of positive substrate samples and a second number of negative substrate samples; and

Based on the first number of positive substrate samples and their production process data, and the second number of negative substrate samples and their production process data, the parameter data reference range of the at least two production equipment parameters is determined.
The method according to claim 3, wherein determining the parameter data reference range of the at least two production equipment parameters comprises:

Using the first number of positive substrate samples and production process data, and the second number of negative substrate samples and production process data to train a production equipment parameter model; and

Using the production equipment parameter model, a parameter data reference range of the at least two production equipment parameters is generated.
3. The method according to claim 2, wherein using the substrate defective distribution pattern of each of the plurality of substrates to generate the defective distribution pattern corresponding to the production process data comprises:

Obtain the respective bad information data of the multiple substrates produced by using the production process data, and generate a substrate bad distribution pattern of each substrate based on the respective bad information data of each substrate; and

Stacking the defective distribution patterns of the substrates of each substrate to obtain the defective distribution patterns corresponding to the production process data.
The method according to claim 2, wherein obtaining at least one reference bad distribution pattern comprises:

Acquire the defective substrate distribution pattern of each defective substrate in the multiple defective substrates in the historical data, wherein the defective substrate distribution pattern shows the substrate coordinate position of the defective point, and the substrate coordinate position of each defective point and its corresponding defective production Equipment parameter correlation; and

Stacking the substrate defective distribution patterns of the plurality of defective substrates in the historical data according to at least one group to obtain the at least one reference defective distribution pattern.
The method according to claim 6, wherein, based on the bad distribution pattern and the at least one reference bad distribution pattern, determining the correlation ranking of each bad production equipment parameter and the bad comprises:

For each defective point in the defective distribution pattern, the substrate coordinate position of the defective point is determined, and the defective production equipment parameter associated with the defective point at the substrate coordinate position in the at least one reference defective distribution pattern is obtained ；

For each identified defective production equipment parameter, determine the number of corresponding defective points; and

According to the number of defective points corresponding to the determined defective production equipment parameters, the correlation between the determined defective production equipment parameters and the defects is sorted.
The method according to claim 6, wherein, based on the bad distribution pattern and the at least one reference bad distribution pattern, determining the correlation ranking of each bad production equipment parameter and the bad comprises:

For each defective point in the defective distribution pattern, determine the substrate coordinate position of the defective point, and obtain the defective production equipment parameters associated with the defective point at the substrate coordinate position in at least one reference defective distribution pattern;

For each identified defective production equipment parameter, determine the number of panels distributed by its corresponding defective point; and

According to the number of panels distributed by the defective points corresponding to the determined defective production equipment parameters, the determined defective production equipment parameters are sorted in correlation.
The method according to claim 7 or 8, wherein obtaining the defective production equipment parameter associated with the defective point at the coordinate position of the substrate in the at least one reference defective distribution pattern comprises:

Displaying the bad distribution pattern and the at least one reference bad distribution pattern;

The parameter data of the production equipment parameters in the production process data corresponding to the defective point at the coordinate position of the substrate on the at least one reference defective distribution pattern is compared with the parameter data reference range, and the defective production equipment exceeding the reference range of the parameter data is obtained The parameter is used as at least one defective production equipment parameter associated with the defective point at the coordinate position of the substrate in the reference defective distribution pattern.
The method of claim 4, wherein:

Extracting multiple sets of production process data of the multiple substrate samples from the distributed storage device.
The method according to claim 10, wherein after the distributed storage device obtains the multiple sets of production process data of the multiple substrate samples, it first performs data preprocessing on each set of production process data to obtain a production conforming to a preset format Process data.
The method according to claim 7, further comprising:

Displaying the defective production equipment parameters in the parameter data of the at least two production equipment parameters that deviate from the parameter data reference range of the at least two production equipment parameters that caused the occurrence of the failure, and/or displaying the sorted and determined defective production equipment parameters.
An electronic device, comprising a memory, a processor, and a computer program stored on the memory and running on the processor, wherein the processor executes the program as described in any one of claims 1 to 12 The method of the cause of the defective substrate is described.
A computer-readable storage medium storing computer instructions for causing the computer to execute the method for determining the cause of substrate failure according to any one of claims 1 to 12.
A system for determining the cause of defective substrates, including:

Distributed storage device, configured to store all production process data of all substrates within a preset time period;

The electronic device according to claim 13; and

The display device is configured to display image data to be displayed obtained from the electronic device.