Predicting Time-to-Failure of Plasma Etching Equipment using Machine\n Learning

2019/04/16 by Anahid Jalali, Jalali, Anahid, Heistracher Clemens +12
Engineering · #Industrial Vision Systems and Defect Detection

paper · pdf · doi:10.48550/arxiv.1904.07686

Abstract

Predicting unscheduled breakdowns of plasma etching equipment can reduce\nmaintenance costs and production losses in the semiconductor industry. However,\nplasma etching is a complex procedure and it is hard to capture all relevant\nequipment properties and behaviors in a single physical model. Machine learning\noffers an alternative for predicting upcoming machine failures based on\nrelevant data points. In this paper, we describe three different machine\nlearning tasks that can be used for that purpose: (i) predicting\nTime-To-Failure (TTF), (ii) predicting health state, and (iii) predicting TTF\nintervals of an equipment. Our results show that trained machine learning\nmodels can outperform benchmarks resembling human judgments in all three tasks.\nThis suggests that machine learning offers a viable alternative to currently\ndeployed plasma etching equipment maintenance strategies and decision making\nprocesses.\n

Related

• In-situ and Non-contact Etch Depth Prediction in Plasma Etching via Machine Learning (ANN & BNN) and Digital Image Colorimetry
• Predictive Maintenance using Machine Learning
• Deep Learning Regression of VLSI Plasma Etch Metrology
• Survival Analysis on Structured Data using Deep Reinforcement Learning
• A two-level machine learning framework for predictive maintenance: comparison of learning formulations