Predictive Analysis and Root Cause Identification of Equipment Failures Using Event Log Patterns: A Focus on EUV/DUV Equipment

被引：0

作者：

Ki, Wanwook ^{[1
]}

Lee, Yongjun ^{[1
]}

Park, Ji Jiyeon ^{[1
]}

Kim, Younghoon ^{[1
]}

Choi, Minjae ^{[1
]}

Park, Jisoo ^{[1
]}

Park, Jaeyong ^{[1
]}

机构：

[1] Samsung Elect, Mechatron Res, Hwaseong, South Korea

来源：

2024 35TH ANNUAL SEMI ADVANCED SEMICONDUCTOR MANUFACTURING CONFERENCE, ASMC | 2024年

关键词：

Prediction; Breakdown; Failure symptom; Event log; Occurrence rarity; EUV; DUV; MODEL;

D O I：

10.1109/ASMC61125.2024.10545492

中图分类号：

T [工业技术];

学科分类号：

08 ;

摘要：

In the rapidly evolving field of semiconductor manufacturing, the productivity and reliability of advanced semiconductor equipment, especially Extreme Ultraviolet (EUV) and Deep Ultraviolet (DUV) lithography systems, are of paramount importance. This research introduces a cutting-edge methodology that employs the analysis of the rarity of event logs to preemptively identify potential failures and their root causes within such equipment. By establishing an optimized threshold that delineates the normal from the rare, this approach effectively flags those rare logs that exceed this threshold as indicators of potential failures. Subsequent analysis of these logs reveals the underlying mechanisms of failure, allowing for timely intervention. The application of this methodology to EUV/DUV equipment has demonstrated a significant increase in predictive accuracy, with a recall of 97.6%, and the potential to realize annual savings of nearly $2 million by reducing equipment downtime and maintenance costs.

引用

页数：5