基於對比式學習之多源晶圓圖搜索於半導體錯誤檢測

Multi-source wafer map retrieval based on contrastive learning for root cause analysis in semiconductor manufacturing

洪緯濬(Wei-Jyun Hong)

指導教授 ： 吳沛遠

國立臺灣大學/電機資訊學院/電信工程學研究所/碩士(2022年)

https://doi.org/10.6342/NTU202203640

摘要

在半導體製造業中，晶圓良率是決定利潤多寡的關鍵因素。由於改善良率涉及分析數百個晶圓製造步驟中的故障，這需要大量的時間和人力，因此先前的研究採用晶圓圖案辨識模型或晶圓圖檢索模型進行故障分析。然而，現有方法嚴重依賴於通過已知的根本原因查找故障，導致對未知根本原因的檢測仍然僅限於透過經驗豐富的工程師手動分析。因此，需要一種檢索模型，通過將良率圖與同一晶圓的完整掃描圖一一比對來直接檢測未知的根本原因，名叫多源晶圓圖檢索模型。儘管現有的多源圖像檢索模型可以彌平來自不同來源中晶圓圖之間的視覺差距，但這些方法大多偏重於尋找語義上相關的樣本（例如相同缺陷類別的樣本），而忽略了多源圖像檢索中應考慮的空間重要性，例如晶圓圖上缺陷圖案的各種尺寸、形狀和位置等等。本研究採用基於對比學習的方法來解決來自不同來源的晶圓圖之間的視覺差距，該方法還考慮了空間信息。本研究還提出了一種混合樣本策略，該策略在來自不同來源的晶圓圖之間應用像素對像素的乘法來顯示共同缺陷位置。我們提出的檢索方法之有效性得到了在線生產晶圓圖數據集測試結果的支持。

關鍵字

根本原因分析 ； 晶圓圖 ； 影像檢索 ； 對比式學習

並列摘要

In semiconductor manufacturing, wafer yield is a key success factor to determine profits. Since yield improvement involves analyzing the failures from hundreds of wafer process steps, which requires a lot of time and manpower, previous studies employ wafer map pattern recognition model or wafer map retrieval model for failure analysis. However, existing methods rely heavily on finding failures through known root cause patterns, resulting in detection of unknown root causes still limited to manual analysis by experienced engineers. A retrieval model is needed to directly detect unknown root causes by matching yield maps to fully scan maps of the same wafer one by one, namely multi-source wafer map retrieval model. Although existing multi source image retrieval models can eliminate the visual gap between wafer maps from different sources, most of these methods focus on finding semantically related samples (e.g. of the same defect category) and ignore the spatial significance that should be considered in multi-source wafer map retrieval, such as various sizes, shapes and locations of defect patterns on wafer maps. This study applies a contrastive learning based approach to address the visual gap between wafer maps from different sources, which also takes spatial information into account. This study also proposes a mixed sample strategy that applies pixelwise multiplication between wafer maps from different sources to indicate common defect locations. The effectiveness of our proposed retrieval approach is supported by testing results on in-line production wafer map datasets.

並列關鍵字

Root cause analysis ； Wafermap ； Image retrieval ； Contrastive learning

參考文獻

[1] T. Chen, S. Kornblith, M. Norouzi, and G. Hinton. A simple framework for contrastive learning of visual representations. In Proceedings of the 37th International Conference on Machine Learning, ICML 2020, volume 119, pages 1597–1607. PMLR, 2020. Virtual Event.

[2] S. Chopra, R. Hadsell, and Y. LeCun. Learning a similarity metric discriminatively, with application to face verification. In IEEE Computer Society Conference on Computer Vision and Pattern Recognition, CVPR 2005, volume 1, pages 539–546, San Diego, California, 2005. IEEE Computer Society.

[3] P. Espadinha-Cruz, R. Godina, and E. M. Rodrigues. A review of data mining applications in semiconductor manufacturing. Processes, 9(2):305, 2021.

[4] J.-B. Grill, F. Strub, F. Altché, C. Tallec, P. Richemond, E. Buchatskaya, C. Doersch, B. Avila Pires, Z. Guo, M. Gheshlaghi Azar, et al. Bootstrap your own latent-a new approach to self-supervised learning. Advances in Neural Information Processing Systems, 33:21271–21284, 2020.

[5] K. He, H. Fan, Y. Wu, S. Xie, and R. Girshick. Momentum contrast for unsupervised visual representation learning. In 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition,CVPR 2020, pages 9726–9735, Seattle, Washington, 2020. Computer Vision Foundation / IEEE.