A novel ontology-based knowledge engineering approach foryield symptom identification in semiconductor manufacturing

  • Authors:

  • Fang-Hsiang Su;Shi-Chung Chang;Chih-Min Fan;Ya-Jung Tsai;Jethro Jheng;Ching-Pin Kao;Chun-Yao Lu

  • Affiliations:

  • Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan, ROC;Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan, ROC;Department of Industrial Engineering and Management, Yuan Ze University, Taoyuan, Taiwan, R.O.C.;Taiwan Semiconductor Manufacturing Co., Hsin-Chu, Taiwan, R.O.C.;Taiwan Semiconductor Manufacturing Co., Hsin-Chu, Taiwan, R.O.C.;Taiwan Semiconductor Manufacturing Co., Hsin-Chu, Taiwan, R.O.C.;Graduate Institute of Industrial Engineering, National Taiwan University, Taipei, Taiwan, ROC

  • Venue:
  • CASE'09 Proceedings of the fifth annual IEEE international conference on Automation science and engineering
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Effective management of knowledge-intensive yield analysis plays a significant role in fast yield ramping for semiconductor manufacturing. Although data analysis platforms with many analysis function tools are available to the industry, there is lack of systematic representation of engineering knowledge for effective extraction and sharing; engineers' identification of situations and analysis purposes and flows are largely in engineers' minds or in disparate forms. In this paper, over the problem domain of fault symptom identification for semiconductor yield analysis, a novel ontology-based modeling framework is first designed for knowledge representations across data, function flow and purpose layers. The ontology model facilitates the knowledge descriptions of an engineer's analysis purpose plan, the application sequences of analysis tools as well as the mapping between a purpose and tool selections. To substantiate the ontology framework with modeling contents, three methods are designed: a Markov chain-based algorithm to extract from engineers' analysis log data their procedures and preferences of tool applications, a tree construction algorithm for engineers' analysis purpose planning, and a graphic symptom capturer for auto-capturing of perceived fault symptoms by engineers. Such designs have been integrated into an engineering data analysis platform that enables engineers' effective extraction, sharing, and reuse of knowledge in situation identification, purpose planning and tool applications.