Method for monitoring stability of semiconductor production and monitoring system

Abstract

Embodiments of the present application provide a method and a system for monitoring stability of semiconductor production. The method comprises: obtaining performance data of semiconductors in different semiconductor processes in a preset time period, wherein different types of machine groups are used in the different semiconductor processes, each type of machine group has different machines, and each machine has different machine chambers; performing first classification processing on the performance data to obtain a number of first failure behaviors of each type of machine group; obtaining an error machine group from the different types of machine groups based on the number of first failure behaviors; performing second classification processing on the performance data of the error machine group to obtain a number of second failure behaviors of each machine chamber; and obtaining and disabling an error machine chamber based on the number of second failure behaviors. Embodiments of the present application can monitor production machines in real time and take measures on abnormal machines to prevent the situation from getting worse.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the field of semiconductor production, and in particular, to a method and system for monitoring stability of semiconductor production. BACKGROUND

[0002] During the production and manufacturing stage of a semiconductor, a plurality of production devices are involved, and different error behaviors occur in each production device. The errors of the devices have a negative impact on the production of the semiconductor.

[0003] The current data system for counting error behaviors of production devices cannot quickly determine whether there is a device tendency problem based on the error data, and the error behaviors of each device are relatively independent and are handled by different personnel. This cannot timely and accurately find the common problems of the devices, thereby causing instability factors in the semiconductor production line and affecting the production quality of the semiconductor. SUMMARY

[0004] Embodiments of the present application provide a method and system for monitoring stability of semiconductor production, which at least helps to monitor the running condition of semiconductor production devices in real time and take quick measures on abnormal devices to prevent the situation from being enlarged.

[0005] According to some embodiments of the present application, the embodiments of the present application provide a method for monitoring stability of semiconductor production, which specifically includes the following steps: obtaining performance data of a semiconductor in different semiconductor processes in a preset time period, different semiconductor processes are performed by different types of machine groups, and each type of machine group has different machines, and each machine has different machine chambers; performing first classification processing on the performance data to obtain a number of first error behaviors of each type of machine group, the number of first error behaviors being obtained based on the performance data of the same type of machine group; obtaining an error machine group from different types of machine groups based on the number of first error behaviors of all machine groups obtained; performing second classification processing on the performance data of the error machine group to obtain a number of second error behaviors of each machine chamber, the number of second error behaviors being obtained based on the performance data of the error machine group; and obtaining and disabling the error machine chamber based on the number of second error behaviors of all machine chambers obtained.

[0006] According to some embodiments of the present application, another aspect of the embodiments of the present application further provides a semiconductor production stability monitoring system, comprising: a collection module, configured to acquire performance data of semiconductors in different semiconductor processes in a preset time period, wherein the different semiconductor processes are performed by different types of machine groups, each type of the machine groups has different machines, and each machine has different machine chambers; a first processing module, configured to perform first classification processing on the performance data, to obtain a number of first failure behaviors of each type of the machine groups, wherein the number of the first failure behaviors is obtained based on the performance data of the same type of the machine groups; a first acquisition module, configured to acquire an error machine group from the different types of the machine groups based on the number of the first failure behaviors of all the machine groups; a second processing module, configured to perform second classification processing on the performance data of the error machine group, to obtain a number of second failure behaviors of each machine chamber, wherein the number of the second failure behaviors is obtained based on the performance data of the error machine group; and a second acquisition module, configured to acquire and disable an error machine chamber based on the number of the second failure behaviors of all the machine chambers.

[0007] The technical scheme provided by the embodiments of the present application has at least the following advantages:

[0008] In the technical scheme of the method for monitoring semiconductor production stability provided by the embodiments of the present application, the performance data of the semiconductors obtained in the preset time period is processed by the first classification processing method, that is, the performance data of the same type of machine groups is summarized, so as to obtain the number of first failure behaviors of each type of machine groups, and the error machine group is found from the different types of machine groups according to the number of first failure behaviors; the performance data of the error machine group is processed by the second classification processing method, that is, the performance data of the error machine group is summarized, so as to obtain the number of second failure behaviors of each machine chamber. By this method, all error machine groups and all error machine chambers can be found more concentratedly, on the one hand, it is easy to judge whether there is a device tendency problem, so as to prevent the error machine group and the error machine chamber from causing adverse effects on the production yield of the semiconductor, on the other hand, the error machine and the error machine chamber can be analyzed and repaired concentratedly, so as to find out the problem in time and find out the fundamental solution. BRIEF DESCRIPTION OF DRAWINGS

[0009] One or more embodiments are illustrated by way of example in the figures that are part of this document, and which illustrate key implementations, and which are not meant to be an exhaustive list of embodiments. Unless otherwise specifically noted, the drawings shown are not necessarily to scale, and the dimensions of the various features can have been arbitrarily scaled or adjusted for clarity.

[0010] Figure 1A flowchart of a method for monitoring semiconductor production equipment according to an embodiment of the present application is provided.

[0011] Figure 2 A flowchart of a method for obtaining an error machine group from different types of machine groups according to an embodiment of the present application is provided.

[0012] Figure 3 A flowchart of a method for obtaining an error machine group from different types of machine groups according to an embodiment of the present application is provided.

[0013] Figure 4 A functional block diagram of a semiconductor production equipment monitoring system according to another embodiment of the present application is provided. DETAILED DESCRIPTION

[0014] As known from the background, a semiconductor production process involves multiple sub-production steps, each of which corresponds to a different production equipment, and each production equipment has the possibility of error. The current production equipment monitoring system cannot quickly locate the error production equipment, and each error behavior is independent of each other, making it difficult to accurately determine the error production equipment and take effective measures in a timely manner.

[0015] The embodiments of the present application provide a method for monitoring the stability of semiconductor production. By classifying the performance data in the semiconductor production process according to different types of production equipment twice, the total number of errors of a specific production equipment within a period of time can be obtained, and the error information corresponding to each production equipment is relatively concentrated, so that whether the production equipment has a device tendency problem can be quickly determined. If so, find out the problem and repair it in time, so as to avoid the device with tendency to cause a larger production accident to the semiconductor production line. The production equipment includes at least one of a machine group, a machine, or a machine chamber.

[0016] The embodiments of the present application will be described in detail below with reference to the accompanying drawings. However, those skilled in the art can understand that in the embodiments of the present application, many technical details are proposed in order to make the reader better understand the present application. However, the technical solutions claimed by the present application can be realized even without these technical details and various changes and modifications based on the following embodiments.

[0017] Figure 1 A flowchart of a method for monitoring the stability of semiconductor production according to an embodiment of the present application is provided.

[0018] Reference Figure 1 The method for monitoring the stability of semiconductor production according to an embodiment of the present application includes the following steps:

[0019] S1: Obtain performance data of different semiconductor processes in a preset time period, different semiconductor processes are performed by different types of machine groups, and each type of machine group has different machines, and each machine has different machine chambers.

[0020] The semiconductor can be a wafer. Specifically, the performance data includes performance data corresponding to abnormal behavior and performance data corresponding to normal operation. The method for obtaining performance data includes: after the wafer completes a corresponding process in a certain machine group or machine chamber, the control system obtains a parameter of the wafer, which can be a film thickness parameter or a pattern feature size parameter; compare the parameter of the wafer with the standard value, if the parameter of the wafer reaches the standard value, it is considered that the performance data corresponding to the machine group or the machine chamber is normal operation, otherwise, it is considered that the performance data corresponding to the machine group or the machine chamber is abnormal behavior.

[0021] The parameter standard value of the wafer refers to the design parameter that the wafer theoretically reaches after completing a corresponding process in a certain machine group or machine chamber.

[0022] It should be noted that the preset time refers to a time period set before the semiconductor data is obtained, and the performance data of the semiconductor is obtained in this time period. The length of the preset time is set according to the frequency of abnormal behavior of the previous machine group, machine and machine chamber. In some embodiments, the preset time is 24 hours, and in other embodiments, the preset time can be 48 hours.

[0023] The abnormal behavior of the machine chamber refers to that the measurement data of the wafer produced and manufactured in the machine chamber does not reach the standard value. However, it should be noted that this may be an abnormal behavior caused by the machine chamber itself, or it may be an abnormal behavior caused by the wafer itself. For example, after the wafer ends the previous production link, the wafer does not reach the qualified thickness due to the problem of the production equipment, and then after the end of the next production link, the thickness of the wafer still does not reach the qualified thickness, at this time, the abnormal behavior generated in the next production link is more likely to be caused by the problem of the wafer itself, rather than the production equipment involved in the next production link.

[0024] S2: Perform first classification processing on the performance data to obtain the number of first abnormal behaviors of each type of machine group, and the number of first abnormal behaviors is obtained based on the performance data of the same type of machine group.

[0025] Among all the obtained performance data, the performance data corresponding to each machine group is independent of each other, and the method of first classification processing is to aggregate the performance data of the same type of machine group. This is conducive to the overall analysis of all abnormal behavior data of the same type of machine group, so as to judge whether the machine group of this type has a tendency error.

[0026] It should be noted that the tendency error of the equipment refers to the persistent working abnormality and performance reduction of the equipment, such as the problem of poor circuit transmission caused by the aging of the circuit.

[0027] In some embodiments, the first classification method is to aggregate the data of the failure behavior in the performance data of the same type of machine group. In other embodiments, the first classification method is to aggregate the data of the failure behavior and the data of the normal operation in the performance data corresponding to the same type of machine group respectively. By comparing the size of the data of the failure behavior and the data of the normal operation, the running status of the machine group can be known.

[0028] The first failure behavior refers to the performance of the wafer manufactured by each type of machine group within a predetermined time not reaching a standard value. Specifically, the performance of the wafer manufactured by a chamber in a machine in a type of machine group does not reach a standard value, that is, the type of machine group is considered to correspond to a first failure behavior.

[0029] In some embodiments, the method for obtaining the number of first failure behaviors includes: obtaining the number of sub-failure behaviors of each machine corresponding to each type of machine group respectively, the number of sub-failure behaviors being obtained based on the performance data of each machine; and obtaining the number of first failure behaviors based on the sum of the number of all sub-failure behaviors corresponding to each type of machine group.

[0030] Specifically, a type of machine group includes a plurality of different machines, the performance data of each machine includes performance data of failure behavior and performance data of normal operation, and the data of failure behavior and the data of normal operation are arranged irregularly. The data of failure behavior is aggregated as the sub-failure behavior of the machine, and the aggregation of the sub-failure behaviors of all machines corresponding to a type of machine group is aggregated as the first failure behavior of the type of machine group.

[0031] In other embodiments, the method for obtaining the number of first failure behaviors can further include: based on the obtained number of sub-failure behaviors, obtaining an error machine from different machines.

[0032] Specifically, a type of machine group includes a plurality of different machines, and the plurality of machines are divided into machines with failure behavior and machines with normal operation. According to the number of sub-failure behaviors of the type of machine group, the machines with failure behavior in different machines are taken as error machines.

[0033] In other embodiments, the method for obtaining the number of first failure behaviors includes: directly obtaining the sum of the number of failure behaviors of all machines corresponding to each type of machine group as the number of first failure behaviors.

[0034] S3: obtaining an error machine group from different types of machine groups based on the number of the first failure behaviors of all the machine groups.

[0035] In some embodiments, referring to Figure 2 The method for obtaining an error machine group from different types of machine groups is: providing a first threshold value; comparing whether the number of the first failure behaviors of each machine group is greater than or equal to the first threshold value; if yes, the machine group with the number of the first failure behaviors greater than or equal to the first threshold value is regarded as an error machine group. It needs to be further explained that the first threshold value is set according to the frequency of the failure behaviors of the machine group in the past, for example, in the past statistical data, the number of the failure behaviors of the machine group within 24 hours is at most 4, and the first threshold value of the machine group is set to 5.

[0036] It needs to be explained that in some embodiments, after obtaining the error machine group, the error machine group can be disabled to avoid the problem of semiconductor scrap caused by continuing to use the error machine group for production.

[0037] In other embodiments, the error machine group can also be selected not to be disabled until the error machine chamber is obtained, and then the error machine group is disabled.

[0038] In some embodiments, it can also include: if the error machine group is obtained, a first alarm information is sent to the relevant personnel in time to remind the relevant personnel to know the running status of the machine group in the semiconductor process, so that the relevant personnel can check and repair the error machine group, and reduce the production loss.

[0039] It needs to be explained that the error machine group is obtained at the same time, and the related information of the error machine group is also stored, such as error time, error number, machine group model and error type, etc. This is beneficial to store the information as historical record for subsequent viewing.

[0040] S4: performing a second classification processing on the performance data of the error machine group, obtaining the number of the second failure behaviors of each machine chamber, and the number of the second failure behaviors is obtained based on the performance data of the error machine group.

[0041] From the obtained performance data of the error machine, the performance data of each machine chamber is independent of each other, and the method of the second classification processing is to summarize the performance data corresponding to each machine chamber. This is beneficial to analyze all the failure behaviors of each machine chamber as a whole, so as to judge whether the machine chamber has a tendency of error.

[0042] In some embodiments, the second classification method is to aggregate the data of the malfunctioning behavior in the performance data of each chamber of the machine tool. In other embodiments, the second classification method is to aggregate the data of the malfunctioning behavior and the data of the normal operation in the performance data of each chamber of the machine tool, respectively. By comparing the size of the data of the malfunctioning behavior and the data of the normal operation, the running status of the chamber of the machine tool can be known.

[0043] In some embodiments, the method of obtaining the number of the first malfunctioning behavior can further include a second classification process of the performance data of the group of the machine tools with errors, which can include a second classification process of the performance data of the machine tools with errors, and obtaining the number of the second malfunctioning behavior of each chamber of the machine tools with errors.

[0044] S5: based on the obtained number of the second malfunctioning behavior of all the chambers of the machine tools, obtaining and disabling the chamber of the machine tool with errors.

[0045] In some embodiments, referring to Figure 3 The method of locating the chamber of the machine tool with errors from different chambers of the machine tools includes: providing a second threshold value, which is smaller than the first threshold value; comparing whether the number of the second malfunctioning behavior of each chamber of the machine tool is greater than or equal to the second threshold value; if yes, the chamber of the machine tool with the number of the second malfunctioning behavior greater than or equal to the second threshold value is taken as the chamber of the machine tool with errors.

[0046] It should be noted that a group of machine tools includes multiple chambers of the machine tools, the first threshold value is the judgment value of the group of machine tools, and the second threshold value is the judgment value of the chamber of the machine tool, so the second threshold value is smaller than the first threshold value.

[0047] It should be further noted that the second threshold value is set according to the frequency of the malfunctioning behavior of the chamber of the machine tool in the past, for example, in the past statistical data, the number of the malfunctioning behavior of the chamber of the machine tool within 24 hours is at most 2, and then the second threshold value of the chamber of the machine tool is set to 3.

[0048] In some embodiments, it further includes: if the chamber of the machine tool with errors is obtained, a second alarm information is sent to the relevant personnel, so as to timely remind the relevant personnel of the running status of the chamber of the machine tool in the semiconductor process, so that the relevant personnel can check and repair the chamber of the machine tool with errors, and reduce the production loss.

[0049] In some embodiments, it further includes: storing the information of the machine tool with errors and the information of the chamber of the machine tool with errors, such as the error time, the error number, the model of the chamber of the machine tool, and the error type, etc. This is beneficial to the storage information as a historical record for subsequent viewing.

[0050] In summary, the number of missing behaviors corresponding to specific machine group or machine chamber is summarized, the relative independence of the missing behaviors is avoided, and it is beneficial to quickly determine whether there is a device tendency. Related personnel can know the device running condition in time through the alarm information, thereby avoiding the lag phenomenon caused by information transmission obstacles and reducing the probability of semiconductor errors. According to the summarized device missing behavior results, a single related personnel detects and maintains the device, which is beneficial to find common problems and find the root solution.

[0051] Correspondingly, another embodiment of the present application also provides a semiconductor production stability monitoring system for executing the method for monitoring the semiconductor production stability in the foregoing embodiments, Figure 4 The function block diagram of the semiconductor production stability monitoring system provided in the present embodiment is as follows, which will be described in detail below Figure 4 The semiconductor production stability monitoring system is described in detail.

[0052] Reference Figure 4 The semiconductor production stability monitoring system comprises a collection module 20, a first processing module 21, a first acquisition module 22, a second processing module 23, a second acquisition module 24, and a third processing module 25.

[0053] The collection module 20 is used to acquire performance data of semiconductors in different semiconductor processes in a preset time period. Different types of machine groups are used for different semiconductor processes, and each type of machine group has different machines, and each machine has different machine chambers.

[0054] The process of collecting data by the collection module 20 is as follows: whether the measurement value of the wafer reaches the standard value is judged. If yes, the corresponding machine chamber runs well, and the data collected by the collection module 20 is that the machine chamber has no missing behavior. Otherwise, the data collected by the collection module 20 is that the machine chamber has missing behavior.

[0055] In some embodiments, the collection module 20 is a collection device on a wafer automated production line. On the one hand, the collection device tests the size performance or physical and chemical performance of the wafer after each production link, and compares the test result with the qualified data of the wafer corresponding to the link; on the other hand, the collection device transmits the comparison result to the control system for the convenience of related personnel to view and analyze.

[0056] It should be noted that the measurement data of the wafer refers to the size data or physical and chemical performance of the wafer measured after the wafer is produced in the machine chamber, such as the height, strength, surface morphology, refractive index, etc. of the wafer.

[0057] It needs to be further explained that not all performance data is processed in the subsequent link, but part of the performance data is obtained in the grabbing time period, and part of the performance data is processed, because the amount of information collected by the collection module 20 is large, and all related information cannot be quickly analyzed. In some embodiments, the grabbing time is 10 min (min, minute).

[0058] The first processing module 21 is used for first classification processing of the performance data, and obtaining the number of first failure behaviors of each type of machine group. The number of first failure behaviors is obtained based on the performance data of the same type of machine group.

[0059] In some embodiments, the first processing module 21 includes a first comparison unit 34 and a first processing unit 35. The first comparison unit 34 is used to provide a first threshold value and compare whether the number of first failure behaviors of each machine group is greater than or equal to the first threshold value. The first processing unit 35 is used to obtain the error machine group based on the comparison result of the first comparison unit.

[0060] The purpose of setting the first threshold value is to compare the number of first failure behaviors of the machine group with the first threshold value to determine whether the machine group has a problem. If so, the first processing unit 35 is used to accurately locate the error machine group. This setting can quickly and conveniently find the error machine group, and can more accurately determine whether there is a device tendency. After the problem is found, the device is checked and repaired by relevant personnel, which facilitates timely discovery of common problems.

[0061] The first acquisition module 22 obtains the error machine group from different types of machine groups based on the number of first failure behaviors of all machine groups obtained.

[0062] In some embodiments, the first acquisition module 22 includes a first acquisition unit 31 and a second acquisition unit 32. The first acquisition unit 31 is used to obtain the number of sub-failure behaviors of the machine corresponding to each type of machine group. The number of sub-failure behaviors is obtained based on the performance data of each machine. The second acquisition unit 32 is used to obtain the first failure behavior based on the sum of the number of all sub-failure behaviors corresponding to each type of machine group based on the number of sub-failure behaviors obtained by the first acquisition unit.

[0063] The second processing module 23 is used for second classification processing of the performance data of the error machine group, and obtaining the number of second failure behaviors of each machine chamber.

[0064] In some embodiments, the second processing module 23 comprises a second comparison unit 36 and a second processing unit 37. The second comparison unit 36 is configured to provide a second threshold value and compare whether the number of second failure behaviors of each chamber of the tool is greater than or equal to the second threshold value. The second processing unit 37 is configured to obtain the tool chamber with error based on the comparison result of the second comparison unit.

[0065] The purpose of setting the second threshold value is similar to the purpose of setting the first threshold value, which will not be repeated here.

[0066] The second obtaining module 24 is configured to obtain and disable the tool chamber with error based on the number of second failure behaviors of all tool chambers obtained.

[0067] In some embodiments, the second obtaining module 24 comprises an obtaining unit 33 configured to perform second classification processing on the performance data of the tool with error, and obtain the number of second failure behaviors of each chamber of the tool with error.

[0068] The third processing module 25 is configured to obtain the tool with error from different tools based on the number of sub-failure behaviors obtained.

[0069] The monitoring system can further comprise an alarm module 26 configured to send first alarm information corresponding to the tool group with error and second alarm information corresponding to the tool chamber with error.

[0070] After receiving the first alarm information or the second alarm information, the semiconductor production system will automatically disable the tool group, tool or tool chamber within the setting, and notify the relevant personnel to make a judgment and take appropriate measures as soon as possible to minimize the risk of on-line products.

[0071] The first alarm information can include alarm siren or sending alarm email to relevant personnel for the tool group with error, and the second alarm information includes alarm siren or sending alarm email to relevant personnel for the tool chamber with error and the corresponding tool group with error, so as to facilitate the relevant personnel to know the running status of the tool group or tool chamber in time, so as to take measures to reduce the occurrence of quality events and prevent the situation from being enlarged, and avoid the time lag phenomenon caused by the relevant personnel checking data and making judgments, and avoid a large number of product problems.

[0072] It should be noted that the relevant personnel can restart the relevant tool group or tool chamber only after the repair of the tool group or tool chamber is completed.

[0073] In summary, the monitoring system provided by the above embodiments collects the failure behaviors of each type of machine group, machine and machine chamber, and when the number of failure behaviors of the equipment reaches a set threshold within a preset time period, the machine group, machine or machine chamber within the set is disabled, and relevant personnel are notified to make a judgment as soon as possible and take corresponding measures to reduce the risk of online products as much as possible.

[0074] Those skilled in the art can understand that the above embodiments are specific embodiments for implementing the present application, and in actual application, various changes can be made in form and details without departing from the spirit and scope of the present application. Any person skilled in the art can make respective changes and modifications without departing from the spirit and scope of the present application, and therefore the protection scope of the present application should be limited by the scope defined in the claims.

Claims

1. A method of monitoring stability of semiconductor production, characterized by, The method comprises the following steps: acquiring performance data of different semiconductor processes in a preset time period, wherein different semiconductor processes are performed by different types of machine groups, and each type of machine group has different machines, and each machine has different machine chambers; performing first classification processing on the performance data to obtain a first failure behavior quantity of each type of machine group, wherein the first failure behavior quantity is obtained based on the performance data of the same type of machine group; based on the obtained first failure behavior quantity of all machine groups, obtaining an error machine group from different types of machine groups; performing second classification processing on the performance data of the error machine group to obtain a second failure behavior quantity of each machine chamber, wherein the second failure behavior quantity is obtained based on the performance data of the error machine group; based on the obtained second failure behavior quantity of all machine chambers, obtaining and disabling the error machine chamber.

2. The method of monitoring stability of semiconductor production according to claim 1, characterized in that, The method for obtaining the first failure behavior quantity comprises the following steps: directly obtaining the sum of the failure behavior quantities of all machines corresponding to each type of machine group as the first failure behavior quantity.

3. The method of monitoring stability of semiconductor production according to claim 1, wherein, The method for obtaining the first failure behavior quantity comprises the following steps: respectively obtaining a sub-failure behavior quantity of each machine corresponding to each type of machine group, wherein the sub-failure behavior quantity is obtained based on the performance data of each machine; based on the sum of all sub-failure behavior quantities corresponding to each type of machine group, obtaining the first failure behavior quantity.

4. The method of monitoring stability of semiconductor production according to claim 3, wherein Further comprising: based on the obtained sub-failure behavior quantity, obtaining an error machine from different machines; the second classification processing on the performance data of the error machine group comprises: performing the second classification processing on the performance data of the error machine to obtain the second failure behavior quantity of each machine chamber in the error machine.

5. The method of monitoring stability of semiconductor production according to claim 1, wherein Further comprising: disabling the error machine group.

6. The method of monitoring stability of semiconductor production according to claim 1, wherein, The method for obtaining an error machine group from different types of machine groups comprises the following steps: providing a first threshold value; comparing whether the first failure behavior quantity of each machine group is greater than or equal to the first threshold value; if yes, regarding the machine group whose first failure behavior quantity is greater than or equal to the first threshold value as the error machine group.

7. The method of monitoring stability of semiconductor production according to claim 6, wherein, Positioning and disabling the error machine chamber comprises the following steps: providing a second threshold value, wherein the second threshold value is smaller than the first threshold value; comparing whether the second failure behavior quantity of each machine chamber is greater than or equal to the second threshold value; if yes, regarding the machine chamber whose second failure behavior quantity is greater than or equal to the second threshold value as the error machine chamber.

8. The method of monitoring stability of semiconductor production according to claim 1, wherein Further comprising: storing the information of the error machine group and the information of the error machine chamber.

9. The method of monitoring stability of semiconductor production according to claim 1, wherein Further comprising: if the error machine group is obtained, issuing a first alarm information; if the error machine chamber is obtained, issuing a second alarm information.

10. A system for monitoring stability of semiconductor production, characterized by, ​ The monitoring system comprises a collecting module, a first processing module, a first obtaining module, a second processing module, and a second obtaining module. The collecting module is configured to acquire performance data of different semiconductor processes within a preset time period, wherein different semiconductor processes are performed by different types of machine groups, and each type of machine group comprises different machines, and each machine comprises different machine chambers. The first processing module is configured to perform first classification processing on the performance data, and obtain a first failure behavior quantity of each type of machine group, wherein the first failure behavior quantity is obtained based on the performance data of the same type of machine group. The first obtaining module is configured to obtain an error machine group from different types of machine groups based on the obtained first failure behavior quantity of all machine groups. The second processing module is configured to perform second classification processing on the performance data of the error machine group, and obtain a second failure behavior quantity of each machine chamber, wherein the second failure behavior quantity is obtained based on the performance data of the error machine group.

11. The system for monitoring stability of semiconductor production according to Claim 10, wherein The second obtaining module is configured to obtain and disable an error machine chamber based on the obtained second failure behavior quantity of all machine chambers. The first obtaining module comprises: A first obtaining unit is configured to obtain a sub-failure behavior quantity of each type of machine group corresponding to the machine, wherein the sub-failure behavior quantity is obtained based on the performance data of each machine.

12. The system for monitoring stability of semiconductor production according to Claim 11, wherein A second obtaining unit is configured to obtain the first failure behavior based on the sum of the sub-failure behavior quantity of each type of machine group corresponding to all sub-failure behavior quantities based on the sub-failure behavior quantity obtained by the first obtaining unit. The monitoring system further comprises a third processing module configured to obtain an error machine from different machines based on the obtained sub-failure behavior quantity.

13. The system for monitoring stability of semiconductor production according to Claim 10, wherein The second obtaining module comprises: An obtaining unit is configured to perform the second classification processing on the performance data of the error machine, and obtain the second failure behavior quantity of each machine chamber in the error machine. The first processing module comprises:

14. The system for monitoring stability of semiconductor production according to Claim 10, wherein A first comparison unit is configured to provide a first threshold value, and compare whether the first failure behavior quantity of each machine group is greater than or equal to the first threshold value. A first processing unit is configured to obtain the error machine group based on the comparison result of the first comparison unit. The second processing module comprises:

15. The system for monitoring stability of semiconductor production according to Claim 10, wherein A second comparison unit is configured to provide a second threshold value, and compare whether the second failure behavior quantity of each machine chamber is greater than or equal to the second threshold value. A second processing unit is configured to obtain the error machine chamber based on the comparison result of the second comparison unit. Further comprising: An alarm module is configured to issue first alarm information corresponding to the error machine group, and issue second alarm information corresponding to the error machine chamber.

Images