簡易檢索 / 詳目顯示

回結果列表
研究生： 楊朝崴
Yang, Chao-Wei
論文名稱： 利用機器學習和光學反射訊號取得高深寬比光阻的關鍵尺寸
Retrieval of critical dimension of photoresist with high aspect ratio using machine learning and optical reflection
指導教授： 張晉愷
Chang, Chin-Kai
學位類別： 碩士
Master
系所名稱： 工學院 - 機械工程學系
Department of Mechanical Engineering
論文出版年： 2023
畢業學年度： 111
語文別： 中文
論文頁數： 88
中文關鍵詞： 高深寬比次波長結構 、反射光譜 、機器學習 、類神經網路
外文關鍵詞： HARSWS, reflection spectra, machine learning, artificial neural network
相關次數： 點閱：176  下載：0 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 製程監控高深寬比結構是半導體產業中的關鍵的技術。傳統的方法使用掃描式或是穿透式電子顯微鏡來監控高深寬比結構的尺寸，但這種方法複雜且耗時。現有半導體的檢測技術中，已經有利用光學的反射訊號來推得電晶體的精確高度和外型，不過由於電晶體的高度約為150 nm，其對應反射訊號和入射光波長之間的關係為單調函數。而高深寬比光阻次波長結構的反射訊號，由於此類光阻具有較大的深度，因此其共振腔反應將會造成不同的波長會對於其反射訊號將會有劇烈的震盪。因此在本研究中，提出了機器學習方法來監控高深寬比次波長結構的尺寸，利用機器學習的演算法來建立其複雜且非單調的反射訊號。在實驗中，使用電子束微影技術製作高深寬比光阻結構，並在此製程中使用正光阻與負光阻製作出形貌互補的結構，增加資料的豐富度，以建構機器學習模型來監控高深寬比光阻結構。為了建立類神經網路模型，分別使用實驗與模擬反射光譜作為輸入和輸出的數據。接著利用有限時域差分法計算不同尺寸的高深寬比光阻結構之模擬反射光譜，並將這些數據做為資料庫使用。一旦從實驗獲得反射光譜數據，就可以使用類神經網路模型和資料庫計算高深寬比光阻結構和資料庫數據之間的相關係數。通過這種方法，可以準確地獲得高深寬比光阻結構的頂部和底部尺寸。

    Process monitoring of high aspect ratio structure with a subwavelength size (HARSWS) is a crucial technology for the semiconductor industry. The traditional method adopts scanning or transmission electron microscopes to monitor the dimension of HARSWS, and it is a complex and time-consuming method. In this study, the machine learning method was proposed to monitor the dimensions of the nanostructure. The electron beam writer was adopted to fabricate the HARSWS. The positive and negative photoresists were used in this lithography process. The two photoresists can provide complementary dimensions to construct machine learning algorithms for the monitoring of HARSWS. To build the artificial neural network (ANN) model, experimental and simulated reflection spectra were used to be input and output variables, respectively. This ANN model can be utilized to generate the simulation-like reflection spectrum from the experimental result. The finite-difference time-domain method was adopted to calculate the simulated reflection spectra for the various dimensions of HARSWS. These simulated data can be the database. Once the reflection data were obtained from the experiment, the ANN model and database can be used to calculate the correlation coefficients for the HARSWS dimensions of the experiment. The top and bottom dimensions of HARSWS can be obtained accurately by this method.

    摘要 I ABSTRACT II 致謝 X 目錄 XI 圖目錄 XIV 表目錄 XVIII 第一章 緒論 1 1.1 研究背景 1 1.2 研究動機 2 第二章 文獻回顧 5 2.1 傳統監測高深寬比結構方法 5 2.1.1 掃描式電子顯微鏡 5 2.1.2 穿透式電子顯微鏡 8 2.1.3 原子力顯微鏡 10 2.2 機器學習 11 2.2.1 機器學習起源 11 2.2.2 機器學習分類 11 2.2.3 類神經網路原理介紹 12 2.2.4 多層感知器訓練原理 17 2.3 機器學習方法預測微奈米結構尺寸案例 23 2.3.1 機器學習方法預測垂直排列二氧化鈦奈米管平均長度 23 2.3.2 機器學習方法預測原子層沉積薄膜厚度 27 2.4 FABRY PEROT共振 33 2.4.1 Fabry Perot 干涉儀 33 2.4.2 高深寬比光阻結構之Fabry Perot共振反射訊號 33 第三章 高深寬比光阻結構製備及機器學習模型建立 35 3.1 監控高深寬比光阻之關鍵尺寸流程設計 35 3.2 機器學習模型建立 35 3.2.1 環境介紹 35 3.2.2 模型選擇 36 3.2.3 資料收集與處理 37 3.2.4 多層感知器網路參數設置 38 3.3 反射光譜模擬 39 3.3.1 模擬軟體FDTD介紹 39 3.3.2 高深寬比結構模擬設定 40 3.3.3 高深寬比結構各尺寸資料庫建立 41 3.4 預測結果與資料庫的相似度比對 41 3.5 實驗儀器原理介紹 43 3.5.1 旋轉塗佈儀 43 3.5.2 電子束微影系統 43 3.5.3 倒置顯微鏡 45 3.5.4 分光光譜儀 45 3.5.5 橢圓偏光儀 47 3.5.6 雙束型聚焦離子束儀 48 3.6 高深寬比光阻結構製備實驗設計 48 3.6.1 高深寬比結構設計 48 3.6.2 實驗流程 49 3.6.3 矽基板清洗 51 3.6.4 光阻塗佈 52 3.6.5 電子束微影曝寫與顯影 54 3.6.6 反射光譜量測 58 3.6.7 SEM影像拍攝 60 第四章 結果與討論 63 4.1 高深寬比光阻結構SEM圖 63 4.2 實驗反射光譜量測結果與模擬之比較 66 4.3 機器學習模型預測結果與誤差 68 4.3.1 Model 1 68 4.3.2 Model 2 75 4.4 資料庫比對結果 82 第五章 結論與未來展望 85 第六章 參考文獻 86

    1. Moore, G.E. Progress in digital integrated electronics. in Electron devices meeting. 1975. Washington, DC.
    2. Bryce, B.A., et al., Silicon nanowire atomic force microscopy probes for high aspect ratio geometries. Applied Physics Letters, 2012. 100(21): p. 213106.
    3. Zhou, W., et al., Fundamentals of scanning electron microscopy (SEM). Scanning Microscopy for Nanotechnology: Techniques and Applications, 2007: p. 1-40.
    4. Gorelick, S., et al., Direct e-beam writing of dense and high aspect ratio nanostructures in thick layers of PMMA for electroplating. Nanotechnology, 2010. 21(29): p. 295303.
    5. Reimer, L., Transmission electron microscopy: physics of image formation and microanalysis. Vol. 36. 2013: Springer.
    6. Williams, D.B., et al., The transmission electron microscope. 1996: Springer.
    7. Goodhew, P.J. and J. Humphreys, Electron microscopy and analysis. 2000: CRC press.
    8. Inkson, B.J., Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) for materials characterization, in Materials characterization using nondestructive evaluation (NDE) methods. 2016, Elsevier. p. 17-43.
    9. Göken, M. and M. Kempf, Microstructural properties of superalloys investigated by nanoindentations in an atomic force microscope. Acta Materialia, 1999. 47(3): p. 1043-1052.
    10. Kempf, M., M. Göken, and H. Vehoff, Nanohardness measurements for studying local mechanical properties of metals. Applied Physics A: Materials Science & Processing, 1998. 66(7).
    11. Binnig, C., Quate, and C. Gerber. Phys. Rev. Lett, 1986. 56(9): p. 930-933.
    12. Butt, H.-J., B. Cappella, and M. Kappl, Force measurements with the atomic force microscope: Technique, interpretation and applications. Surface science reports, 2005. 59(1-6): p. 1-152.
    13. Knittel, P., et al., Focused ion beam-assisted fabrication of soft high-aspect ratio silicon nanowire atomic force microscopy probes. Ultramicroscopy, 2017. 179: p. 24-32.
    14. Rosenblatt, F., The perceptron, a perceiving and recognizing automaton Project Para. 1957: Cornell Aeronautical Laboratory.
    15. Sharma, S., S. Sharma, and A. Athaiya, Activation functions in neural networks. Towards Data Sci, 2017. 6(12): p. 310-316.
    16. Glorot, X. and Y. Bengio. Understanding the difficulty of training deep feedforward neural networks. in Proceedings of the thirteenth international conference on artificial intelligence and statistics. 2010. JMLR Workshop and Conference Proceedings.
    17. LeCun, Y., et al., Efficient backprop, in Neural networks: Tricks of the trade. 2002, Springer. p. 9-50.
    18. Nair, V. and G.E. Hinton. Rectified linear units improve restricted boltzmann machines. in Proceedings of the 27th international conference on machine learning (ICML-10). 2010.
    19. Maas, A.L., A.Y. Hannun, and A.Y. Ng. Rectifier nonlinearities improve neural network acoustic models. in Proc. icml. 2013. Atlanta, Georgia, USA.
    20. 李弘毅. 機器學習課程講義. Available from: http://speech.ee.ntu.edu.tw/~tlkagk/courses_ML17_2.html.
    21. Rumelhart, D.E., G.E. Hinton, and R.J. Williams, Learning representations by back-propagating errors. nature, 1986. 323(6088): p. 533-536.
    22. 葉怡成, 類神經網路模式應用與實作. 第八版 ed. 2004.
    23. Caro-Gutiérrez, J., et al., Machine learning for predicting the average length of vertically aligned TiO2 nanotubes. AIP Advances, 2020. 10(7): p. 075116.
    24. Arunachalam, A., et al., Machine learning approach to thickness prediction from in situ spectroscopic ellipsometry data for atomic layer deposition processes. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, 2022. 40(1): p. 012405.
    25. Dillon, A., et al., Surface chemistry of Al2O3 deposition using Al (CH3) 3 and H2O in a binary reaction sequence. Surface Science, 1995. 322(1-3): p. 230-242.
    26. Park, S., et al., Atomic layer deposition of Ru/RuO2 thin films studied by in situ infrared spectroscopy. Chemistry of Materials, 2010. 22(17): p. 4867-4878.
    27. Bernal-Ramos, K., et al., Atomic layer deposition of cobalt silicide thin films studied by in Situ infrared spectroscopy. Chemistry of Materials, 2015. 27(14): p. 4943-4949.
    28. Kaufman-Osborn, T., et al., Atomic imaging and modeling of passivation, functionalization, and atomic layer deposition nucleation of the SiGe (001) surface via H2O2 (g) and trimethylaluminum dosing. Surface science, 2014. 630: p. 273-279.
    29. Kaufman-Osborn, T., E.A. Chagarov, and A.C. Kummel, Atomic imaging and modeling of H2O2 (g) surface passivation, functionalization, and atomic layer deposition nucleation on the Ge (100) surface. The Journal of Chemical Physics, 2014. 140(20): p. 204708.
    30. Langereis, E., et al., In situ spectroscopic ellipsometry as a versatile tool for studying atomic layer deposition. Journal of Physics D: Applied Physics, 2009. 42(7): p. 073001.
    31. Langereis, E., et al., In situ spectroscopic ellipsometry study on the growth of ultrathin TiN films by plasma-assisted atomic layer deposition. Journal of applied Physics, 2006. 100(2): p. 023534.
    32. 吳思緯, 以Fabry Perot Etalon 做雷射高斯光束發散角量測之研究, in 光機電工程研究所. 2012, 國立中央大學: 桃園縣. p. 86.
    33. Okugi, T. and K. Okubo, High-speed FDTD calculation method specialized for automotive radar analysis. IEICE Communications Express, 2022. 11(4): p. 165-170.
    34. Lee, K.-H., et al., Modeling and investigation of a geometrically complex UWB GPR antenna using FDTD. IEEE Transactions on Antennas and Propagation, 2004. 52(8): p. 1983-1991.
    35. Yee, K., Numerical solution of initial boundary value problems involving Maxwell's equations in isotropic media. IEEE Transactions on antennas and propagation, 1966. 14(3): p. 302-307.
    36. Taylor, R., Interpretation of the correlation coefficient: a basic review. Journal of diagnostic medical sonography, 1990. 6(1): p. 35-39.
    37. 成功大學微奈米中心. 儀器列表. Available from: http://cfc2021.cfc.ncku.edu.tw/cmnst_e_system/index.php/facility_list/list.
    38. Inc, T.F.S. Thermo Fisher Scientific 官網. Available from: https://www.thermofisher.com/tw/zt/home.html.
    39. microresisttechnology. micro resist technology官網. Available from: https://www.microresist.de/en/.

    下載圖示
    2026-08-31公開
    QR CODE