Semiconductor Processing Alarm Analysis Using Near-Threshold Sensor Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor manufacturing, as devices miniaturize, managing the increasing amount of data from substrate processing apparatuses becomes challenging, leading to difficulties in identifying the cause of abnormalities and reducing downtime, which hampers productivity and production efficiency.
Innovation Solution
A technique that outputs alarms based on sensor information, acquires and compares apparatus data, and identifies specific data closest to threshold values to determine the cause of alarms, facilitating quicker troubleshooting and reducing downtime.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data management is performed in more detail to handle increasing data amounts, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The management apparatus is segmented into multiple processing units: a determination unit that compares sensor data with threshold values, an identification unit that identifies abnormality causes when thresholds are exceeded, and a display unit that presents results. This segmentation allows detailed data management while distributing complexity across specialized functional modules.
Solution Approach 2:
The management apparatus acts as an intermediary between the substrate processing apparatus and users/operators. It automatically collects sensor data, performs threshold comparisons, identifies abnormality causes, and displays results, thereby managing detailed data without requiring users to directly handle the complexity of data analysis.
2Manufacturing precision
If manual troubleshooting is performed to identify abnormality causes, then manufacturing precision is maintained, but loss of time increases
Solution Approach 1:
The management apparatus performs preliminary actions by continuously monitoring sensor data and comparing it with threshold values in advance. When abnormality causes are identified through automatic comparison and analysis, the system is prepared with diagnostic information before actual troubleshooting is needed, thereby maintaining accuracy while reducing response time.
Solution Approach 2:
The management apparatus performs self-service by automatically identifying abnormality causes without requiring skilled personnel intervention. The determination unit and identification unit autonomously analyze sensor data, compare values, and generate diagnostic results, freeing operators from time-consuming manual troubleshooting while maintaining identification accuracy.
3Productivity
If automated monitoring is implemented to reduce reliance on skilled personnel, then productivity is improved, but device complexity increases
Solution Approach 1:
The management apparatus is designed with multi-functionality to justify its complexity: it collects sensor data, compares values against thresholds, identifies abnormality causes, and displays results. By consolidating these multiple functions into a single automated system, productivity improves through reduced reliance on skilled personnel while the complexity is centralized and managed within one universal platform.
Data Source
AI summary
According to one aspect of a technique the present disclosure, there is provided a method of identifying a cause of an abnormality, including: (a) outputting an alarm indicating the failure detected based on sensor information; (b) acquiring a plurality of apparatus data comprising a plurality piece of the sensor information related to the alarm; (c) comparing each of the apparatus data with a threshold value; and (d) when none of the apparatus data acquired in (b) exceed the threshold value according to a comparison result in (c), identifying a specific apparatus data among the plurality of apparatus data closest to the threshold value as a factor of generating the alarm.