針對先進製程中系統性缺陷以及微小延遲缺陷並考慮實體及時序資訊之診斷技術

Physical-aware and Timing-aware Diagnosis for Systematic Defects and Small Delay Defects in Advanced Technology

陳柏瑞(Po-Juei Chen)

指導教授 ： 李建模

國立臺灣大學/電機資訊學院/電子工程學研究所/博士(2014年)

https://doi.org/10.6342/NTU.2014.02529

摘要

在先進製程中，由於晶片中電晶體密度的增加以及操作頻率的上升，使得系統性缺陷 (systematic defects) 及微小延遲缺陷 (small delay defects) 成為良率及可靠性改善中的重要挑戰。 為了驗證可能造成缺陷的原因，實體錯誤分析(physical failure analysis) 是最普遍被使用的技術。 然而，由於先進設計電路的複雜度快速增加，實體錯誤分析所需要的時間不斷上升。 因此，非常需要準確的診斷技術來妥善選擇進行實體錯誤分析的晶片。
針對系統性缺陷，我們提出了診斷技術來鑑別有可能造成良率下降的致命實體特徵(culprit physical features)。 首先，我們從晶片佈局抽取容易造成缺陷的實體特徵，每個特徵分別與製程中不同發生缺陷的原因有關，例如:壓力，不完美的平面化，與金屬線和通孔(via)間不足的接觸面積等。 接著，對於每顆有測試錯誤的晶片，找出可疑的缺陷實體位置。 在這階段我們提出考慮實體佈局資訊與實體特徵並能處理多重缺陷(multiple defects)的診斷技術。 最後，我們使用變異數分析(analysis of variance)的統計技術來識別出可能的致命實體特徵。 我們所提出的技術在55奈米的製程中用來診斷209顆在測試中有錯誤的晶片，並成功地鑑別了一個可能造成良率下降的致命實體特徵。 根據此特徵而選擇的四顆晶片則經由實體錯誤分析證實了有系統性開路缺陷(systematic open defects)的存在。
針對微小延遲缺陷，我們則提出了考慮時序資訊的診斷技術，並使用高於操作頻率的測試技術(faster-than-at-speed test)。 我們使用了時序分析得到的上下界資訊(timing upper bound and lower bound)來改善診斷結果。 我們所提出的技術在先進製程中用來測試及診斷25顆先進工業界晶片。 其中一顆晶片的診斷結果並經由實體錯誤分析驗證確認有微小延遲缺陷的存在。

關鍵字

系統性缺陷 ； 微小延遲缺陷 ； 致命實體特徵 ； 考慮實體資訊之診斷 ； 考慮時序資訊之診斷 ； 多重缺陷診斷 ； 統計分析診斷

並列摘要

Systematic defects and small delay defects (SDD) have become key challenges of yield and reliability due to shrinking geometry and increasing frequency in advanced technology. For yield and reliability improvement, physical failure analysis (PFA) is the most widely-used method to understand defect mechanisms. However, PFA is usually performed on a small part of failing dies because it is very time-consuming and expensive. Hence, a diagnosis technique which can correctly identify systematic defects and SDD to guide the selection of PFA dies is very much needed.
For systematic defect diagnosis, identification of culprit physical features that are responsible for yield loss is important for both yield enhancement and design-for-manufacturability (DFM) rule evaluation. Culprit physical features are certain defect-prone layout characteristic associated with various systematic defect mechanisms, such as stress, imperfect planarization, and other complex design-process interactions. To identify culprit physical features, we propose a systematic defect diagnosis technique, which considers physical feature in diagnosis of failing dies and then performs statistical analysis. To diagnose defects with complex failing behavior and multiple defects in the same die, we propose two physical-aware diagnosis techniques: physical feature-based diagnosis (PF-D) and multiple defect physical-aware diagnosis (MD-PhD). To cope with noise from random defects, a statistical technique

並列關鍵字

systematic defects ； small delay defects ； culprit physical features ； physical-aware diagnosis ； timing-aware diagnosis ； multiple defect diagnosis ； statistical diagnosis

參考文獻

[Acken 1992] J. M. Acken, and S. Millman, "Fault model evolution for diagnosis: Accuracy vs precision," Custom Integrated Circuits Conference, pp. 13.4. 1-13.4. 4, 1992.

[Ahmed 2006a] N. Ahmed, M. Tehranipoor, and V. Jayaram, "Timing-based delay test for screening small delay defects," Design Automation Conference, pp. 320-325, 2006a.

[Ahmed 2006b] N. Ahmed, M. Tehranipoor, and V. Jayaram, "A novel framework for faster-than-at-speed delay test considering IR-drop effects," International Conference on Computer-Aided Design, pp. 198-203, 2006b.

[Balachandran 1999] H. Balachandran, J. Parker, D. Shupp, S. Butler, K. M. Butler, et al., "Correlation of logical failures to a suspect process step," International Test Conference, pp. 458-466, 1999.

[Bartenstein 2001] T. Bartenstein, D. Sliwinski, D. Heaberlin, and L. Huisman, "Diagnosing combinational logic designs using the single location at-a-time (SLAT) paradigm," International Test Conference, pp. 287-287, 2001.