CN-121996689-A - Quick query method, device, medium, program product and terminal for machine defect trend and historical actions

CN121996689ACN 121996689 ACN121996689 ACN 121996689ACN-121996689-A

Abstract

The application provides a rapid inquiring method, a rapid inquiring device, a rapid inquiring medium, a rapid inquiring program product and a rapid inquiring terminal for machine defect trend and historical actions, provides integrated inquiring and analyzing functions, and remarkably improves machine management and fault coping capacity. The user can easily inquire the detailed action and state of the machine, and acquire the closed-loop result of the case to evaluate the effect of the improvement measure, and all the information is presented through an intuitive chart. By integrating the machine state chart and defect data analysis, a user can comprehensively know the history operation condition of the machine, identify abnormal reasons and effective countermeasures, and provide basis for quick response, so that the influence quantity of wafers is reduced, and the yield fluctuation is controlled. Real-time and historical data query is supported, the working efficiency is greatly improved, and the problems of slow speed and complex process of manual query are avoided. Meanwhile, the database comprehensively records detailed action information of the machine, overcomes data incompleteness, and enables a user to analyze defect distribution and directivity problems in a specific time period.

Inventors

LIANG SHIFA
DU LI
HUANG SHENGJING
DING YONGJIE
LI ZHENJUN
ZENG PING

Assignees

华润微电子（重庆）有限公司

Dates

Publication Date: 20260508
Application Date: 20241106

Claims (10)

1. A quick query method for machine defect trend and historical actions is characterized by comprising the following steps: In the process of machine execution, performing defect detection on a wafer placed on a machine to generate defect data, and simultaneously recording operation actions of the machine to generate action data; When the defect data is detected to exceed the machine table abnormality threshold, preprocessing operation is carried out on the action data, an action time chart is generated according to the action data after the preprocessing operation, a defect time chart is generated based on the defect data, and a machine table state chart is generated based on the action time chart and the defect time chart.
2. The rapid query method of machine defect trend and historical motion of claim 1, wherein the act of performing a preprocessing operation on the motion data comprises generating a corresponding conversion file based on the motion data.
3. The rapid query method for machine defect trends and historical actions according to claim 2, wherein generating the corresponding conversion file based on the action data comprises: Performing a data cleaning operation and a time normalization operation on the action data to generate pre-processed action data; Assigning a unique label to each different type of action in the pre-processing action data to generate an action label corresponding to each action type; And constructing a conversion data structure based on the action labels and the action types which are in one-to-one correspondence to generate the conversion file.
4. The rapid query method of claim 2, wherein generating a machine state map based on the action time map and the defect time map comprises: obtaining a defect state and a corresponding time stamp from a defect time chart, and obtaining an operation action and a time stamp from an action time chart; Performing a data merge operation based on the time stamp of the defect time map and the time stamp of the action time map to generate machine state data including the time stamp, the defect state, and the action; And generating a machine state diagram for representing the trend of machine motion change and defect state change based on the machine state data.
5. The rapid query method for machine defect trends and historical actions according to claim 4, wherein the machine state diagram is generated based on the action time diagram and the defect time diagram, and further comprising: acquiring a real-time defect state and a corresponding time stamp of the machine; Extracting an optimized operation action corresponding to the current defect state from the machine state diagram based on the real-time defect state and the corresponding timestamp; and converting the optimized operation action into a machine executable instruction, sending the machine executable instruction to the machine, and monitoring and analyzing the execution result.
6. The method of claim 1, wherein performing defect inspection on a wafer disposed on a tool to generate defect data comprises: Image acquisition is carried out on a wafer placed on a machine table so as to generate a wafer image file; based on an image recognition algorithm, recognizing in the wafer image file to generate a defect image marked with a plurality of different types; and judging each identified defect image based on a preset defect threshold parameter to judge whether the current defect image belongs to an abnormal condition or not, and generating defect data according to the defect image belonging to the abnormal condition.
7. The utility model provides a quick inquiry unit of board defect trend and historical action which characterized in that includes: The data acquisition and preprocessing module is used for carrying out defect detection on a wafer placed on the machine in the machine execution process so as to generate defect data, and recording the operation action of the machine so as to generate action data; The machine state visualization query module is used for executing preprocessing operation on the action data when the defect data is detected to exceed a machine abnormality threshold, generating an action time chart according to the action data after the preprocessing operation, generating a defect time chart based on the defect data, and generating a machine state chart based on the action time chart and the defect time chart.
8. A computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the machine defect trend and history action fast query method of any one of claims 1 to 6.
9. A computer program product, characterized in that the computer program product comprises computer program code which, when run on a computer, causes the computer to implement the machine defect trend and history action fast query method according to any one of claims 1 to 6.
10. An electronic terminal comprising a memory, a processor and a computer program stored on the memory, wherein the processor executes the computer program to implement the machine defect trend and historical motion fast query method of any one of claims 1 to 6.

Description

Quick query method, device, medium, program product and terminal for machine defect trend and historical actions Technical Field The present application relates to the field of semiconductor device manufacturing, and in particular, to a method, an apparatus, a medium, a program product, and a terminal for quickly querying a machine defect trend and a history action. Background In modern semiconductor manufacturing, machines (equipment) are core components in the production process, and their operating state and performance directly affect the quality and production efficiency of the product. With the increasing complexity of the production process, the machine generates a lot of data during the operation process, including detailed Action records (actions) and Defect trends (defects). These data are critical to analyzing the operating state of the machine, identifying potential problems, and optimizing the production process. However, the current machine state query system has a plurality of defects. The status query results of each machine are often displayed in a complex manner, and operators need to screen out useful data from a large amount of information so as to analyze defect conditions and actions. This process is complex and time consuming, often requiring confirmation by manual means such as telephone, increasing the uncertainty and risk of error in the information transfer. In addition, the defect trend and the action history of the machine are usually stored independently, and an effective integration and analysis tool is lacking, so that the overall state of the machine cannot be obtained by one key. In Case-like (Case) tracking, the prior art lacks a machine action analysis function for a specific Case, nor does it summarize actions taken by the same machine or model in similar cases. This makes it necessary for the operator to perform a plurality of manual operations and judgment when handling an abnormal event, and the reaction time is prolonged, thereby affecting the production efficiency and the product quality. Disclosure of Invention In view of the above-mentioned drawbacks of the prior art, the present application is directed to a method, apparatus, medium, program product and terminal for quickly inquiring a machine defect trend and a historical action, which are used for solving the problems that an operator needs to spend a lot of time for data screening and manual confirmation due to the complicated and insufficient integration of a machine state inquiring system, thereby affecting the production efficiency and the product quality. In order to achieve the above and other related objects, a first aspect of the present application provides a method for quickly querying a trend and historical actions of a machine defect, where the method includes performing defect detection on a wafer placed on a machine to generate defect data while recording operation actions of the machine to generate action data, performing preprocessing operation on the action data when the defect data is detected to exceed a machine abnormality threshold, generating an action time map according to the action data after the preprocessing operation, generating a defect time map based on the defect data, and generating a machine state map based on the action time map and the defect time map. In some embodiments of the first aspect of the present application, the process of performing a preprocessing operation on the action data includes generating a corresponding conversion file based on the action data. In some embodiments of the first aspect of the present application, generating a corresponding conversion file based on the action data includes performing a data cleansing operation and a time normalization operation on the action data to generate pre-processed action data, assigning a unique label to each different type of action in the pre-processed action data to generate an action label corresponding to each action type, and building a conversion data structure based on the one-to-one action labels and action types to generate the conversion file. In some embodiments of the first aspect of the present application, the process of generating a machine state map based on the action time map and the defect time map includes obtaining a defect state and a corresponding timestamp from the defect time map, obtaining an operation action and a timestamp from the action time map, performing a data merging operation based on the timestamp of the defect time map and the timestamp of the action time map to generate machine state data including the timestamp, the defect state and the operation action, and generating a machine state map for characterizing a trend of machine action change and defect state change based on the machine state data. In some embodiments of the first aspect of the present application, after generating the machine state diagram based on the action time diagram and the defect time diagram, the method f