考慮光學之實體層系統性短路缺陷診斷

Abstract

在先進奈米製成技術下，利用大量診斷來找出系統性缺陷來提高良率是一件很重要的課題。隨著設計越來越複雜與製成越來越先進，光學微影的影響對於良率有越來越大的影響，儘管布局沒有違反設計規則，在經過光學模擬之後，還是會有機會發現造成光學微影上的缺陷，過去大量診斷技術並未將光學模擬加入考慮。因此，本篇論文提出了一個考慮光學微影的大量診斷技術。在診斷的過程中考慮光學微影，並在最後利用統計分析方法 (Z-test) 找出是否有系統性的短路缺陷 (Layer of Systematic Defect for bridging fault) 。實驗結果在使用3次偵測測試向量時，在單一與多重缺陷皆成功找出系統性缺陷，實驗結果也指出利用多次偵測測試向量有助於提高診斷系統性缺陷的精準度。

Volume diagnosis to identify systematic defects is important for yield learning in modern nanometer technologies.The impact of lithography has become more and more critical to the yield because of increasing design complexity and shrinking technology nodes.Existing volume diagnosis techniques do not consider the lithography information. Therefore, this thesis proposes a volume diagnosis technique that considers the lithography information and uses statistical analysis, Z-test, to identify the layer of systematic defects (LSD) for bridging faults.The experiments successfully diagnoseall LSDs for single bridging fault as well as multiple bridging faultswhen three-detect test patterns are applied. It also shows that N-detect test patterns help to improve the diagnosis results.