Knowledge engineering of analysis tool application processes for yield symptom identification

Effective management of knowledge-intensive yield analysis plays a significant role in fast yield ramping. Over the problem domain of fault symptom identification in semiconductor yield anlaysis, three mechanisms are designed in this paper to extract the knowledge of engineers. The mechanisms include Unified Resource Model-based purpose and tool mapping for linking engineers´ analysis purposes to analysis tools, Markov chain-based knowledge extraction for reusing anlaysis tool procedure knowledge, and Graphic symptom capturer for auto-capturing perceived fault symptoms of engineers. Such designs are integrated into a service oriented architecture-based engineering data analysis platform to demonstrate their feasibility and potential for effective management of yield analysis knowledge.