化學機械研磨(Chemical-Mechanical Polishing)普遍應用於半導體先進製程，主要製程目的為達成全域平坦化(Global planarization)。而在製程尺度縮至奈米級別的現今，CMP製程需要控制之平坦度也隨之越精確。多種因子的影響之下，實際的CMP系統非常複雜並且以傳統模型預測非常具有挑戰性。
　　近年各研究團隊將深度學習技術引入，提出以機器學習之方法進行CMP之預測，與傳統模型相較，機器學習在許多方面皆有較佳之預測表現。然而機器學習方法有多種演算法與架構，本研究將列舉數種機器學習模型並且以貝式最佳化優化之，發現因溫度殘留效應，BiLSTM(Bi-directional Long Short-Term Memory)演算法對於CMP之資料集在各項指標具有壓倒性之表現。更進一步將此模型導入並改善個案CMP製程之平坦度能力與產出，進一步提升晶圓良率與品質。

Chemical-Mechanical Polishing (CMP) is a general manufacturing process used to achieve global planarization in the semiconductor industry. Nowadays, the smaller the semiconductor’s design rules are, the more precise the CMP process should be. Since the actual CMP system is very complicated and dependent on various conditions, the efficient prediction of flatness results using traditional models becomes more challenging.
In the past few years, machine learning skills have been introduced to CMP; many algorithms and structures have been developed accordingly. Compared with traditional models, the machine learning one demonstrates superior performance in every aspect. Also, many algorithms have been proposed. In this study, we evaluated various Neural-Network algorithms and modified them via Bayesian optimization. Since the residual temperature affects CMP systems, the Bi-directional Long Short-Term Memory (BiLSTM) algorithm has an overwhelming advantage over other evaluated models. In this study, optimal hyperparameters have been found. Furthermore, the proposed model has been implemented on-site to significantly improve the CMP flatness capability and throughput for a specific case.

摘要 i
Abstract ii
Acknowledgement iii
List of Tables vi
List of Figures vii
1 Introduction 1
2 Literature Review 5
2.1 The CMP Model’s Evolution 5
2.2 Machine Learning 8
2.3 Hyper-Parameter Optimization 13
3 Methodology 16
3.1 Data Description 16
3.2 Recurrent Neural Network Algorithms 22
3.3 Bayesian Optimization 26
3.4 Performance Evaluation 31
4 Experiments and Results 33
4.1 The Experiment of Deep Learning Models 33
4.2 Model Compression and Results 38
4.3 Implementation and Results 40
5 Conclusion 44
References 46

Albawi, S., Mohammed, T. A., and Al-Zawi, S. (2017). Understanding of a Convolutional Neural Network. In 2017 International Conference on Engineering and Technology (ICET), pages 1–6. IEEE.
Babu, S. (2016). Advances in Chemical Mechanical Planarization (CMP). Woodhead Publishing.
Bao, H. and Chen, L. (2020). A CNN-Based CMP Planarization Model Considering LDE Effect. IEEE Transactions on Components, Packaging and Manufacturing Technology, 10(4):723–729.
Bates, S. P. (2000). Silicon Wafer Processing. Industry Initiatives for Science and Math Education (IISME), San Jose, CA.
Botchkarev, A. (2018). Performance Metrics (Error Measures) in Machine Learning Regression, Forecasting and Prognostics: Properties and Typology. arXiv preprint arXiv:1809.03006.
Chung, J., Gulcehre, C., Cho, K., and Bengio, Y. (2014). Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling. arXiv preprint arXiv:1412.3555.
Cook, L. M. (1990). Chemical Processes in Glass Polishing. Journal of Non-Crystalline Solids, 120(1-3):152–171.
de Roover, D., Emami-Naeini, A., and Ebert, J. (2004). Model-Based Control for Chemical-Mechanical Planarization (CMP). In Proceedings of the 2004 American Control Conference, volume 5, pages 3922–3929. IEEE.
Dewancker, I., McCourt, M., and Clark, S. (2015). Bayesian Optimization Primer. SigOpt.
Du, X., Xu, H., and Zhu, F. (2021). Understanding the Effect of Hyperparameter Optimization on Machine Learning Models for Structure Design Problems. Computer Aided Design, 135:103013.
Fan, R. and Liu, G. (2018). CNN-Based Audio Front End Processing on Speech Recognition. In 2018 International Conference on Audio, Language and Image Processing (ICALIP), pages 349–354. IEEE.
Feurer, M. and Hutter, F. (2019). Hyperparameter Optimization. In Automated Machine Learning, pages 3–33. Springer, Cham.
Francisco, L., Mao, R., Katakamsetty, U., Verma, P., and Pack, R. (2019). Multilayer CMP Hotspot Modeling Through Deep Learning. In Design Process Technology Cooptimization for Manufacturability XIII, volume 10962, pages 202–213. SPIE.
Fu, G., Chandra, A., Guha, S., and Subhash, G. (2001). A Plasticity-Based Model of Material Removal in Chemical-Mechanical Polishing (CMP). IEEE Transactions on Semiconductor Manufacturing, 14(4):406–417.
Ghulghazaryan, R. and Wilson, J. (2017). Application of Machine Learning and Neural Networks for Generation of Pre-CMP Profiles of Advanced Deposition Processes for CMP Modeling. In ICPT 2017; International Conference on Planarization/CMP Technology, pages 1–6. VDE.
Graves, A. and Schmidhuber, J. (2005). Framewise Phoneme Classification with Bidirectional LSTM and Other Neural Network Architectures. Neural Networks, 18(5-6):602–610.
Hochreiter, S. and Schmidhuber, J. (1997). Long Short-Term Memory. Neural Computation, 9(8):1735–1780.
Jayalakshmi, T. and Santhakumaran, A. (2011). Statistical Normalization And Back Propagation for Classification. International Journal of Computer Theory and Engineering, 3(1):1793–8201.
LeCun, Y., Bengio, Y., and Hinton, G. (2015). Deep Learning. Nature, 521(7553):436–444.
Li, Z., Wu, D., and Yu, T. (2019). Prediction of Material Removal Rate for Chemical Mechanical Planarization Using Decision Tree-Based Ensemble Learning. Journal of Manufacturing Science and Engineering, 141(3).
Lipton, Z. C., Berkowitz, J., and Elkan, C. (2015). A Critical Review of Recurrent Neural Networks for Sequence Learning. arXiv preprint arXiv:1506.00019.
Liu, C. W., Dai, B. T., Tseng, W. T., and Yeh, C. F. (1996). Modeling of the Wear Mechanism during Chemical-Mechanical Polishing. Journal of the Electrochemical Society, 143(2):716.
Lujan-Moreno, G. A., Howard, P. R., Rojas, O. G., and Montgomery, D. C. (2018). Design of Experiments and Response Surface Methodology to Tune Machine Learning Hyperparameters, with a Random Forest Case-Study. Expert Systems with Applications, 109:195–205.
Maind, S. B., Wankar, P., et al. (2014). Research Paper on Basic of Artificial Neural Network. International Journal on Recent and Innovation Trends in Computing and Communication, 2(1):96–100.
McCulloch, W. S. and Pitts, W. (1943). A Logical Calculus of the Ideas Immanent in Nervous Activity. The Bulletin of Mathematical Biophysics, 5(4):115–133.
McGrath, J. and Davis, C. (2004). Polishing Pad Surface Characterisation in Chemical Mechanical Planarisation. Journal of Materials Processing Technology, 153:666–673.
Munem, M., Bashar, T. M. R., Roni, M. H., Shahriar, M., Shawkat, T. B., and Rahaman, H. (2020). Electric Power Load Forecasting Based on Multivariate LSTM Neural Network Using Bayesian Optimization. In 2020 IEEE Electric Power and Energy Conference (EPEC), pages 1–6. IEEE.
Oliver, M. R. (2004). Chemical-Mechanical Planarization of Semiconductor Materials, volume 69. Springer Science & Business Media.
Preston, F. W. (1927). The Theory and Design of Plate Glass Polishing Machines. Journal of Glass Technology, 11(44):214–256.
Rasamoelina, A. D., Adjailia, F., and Sincˇak, P. (2020). A Review of Activation Function ´ for Artificial Neural Network. In 2020 IEEE 18th World Symposium on Applied Machine Intelligence and Informatics (SAMI), pages 281–286. IEEE.
Ruder, S. (2016). An Overview of Gradient Descent Optimization Algorithms. arXiv preprint arXiv:1609.04747.
Schmid, P. E. (1981). Optical Absorption in Heavily Doped Silicon. Physical Review B, 23(10):5531.
SEMI (2007). Test Method for Measuring Flatness, Thickness, and Total Thickness Variation on Silicon Wafers by Automatic Non-Contact Method. Semiconductor Equipment and Materials International, San Jose, CA, USA, SEMI Standard MF1530–0707.
Seo, J. (2021). A Review on Chemical and Mechanical Phenomena at the Wafer Interface during Chemical Mechanical Planarization. Journal of Materials Research, 36(1):235–257.
Shahriari, B., Swersky, K., Wang, Z., Adams, R. P., and De Freitas, N. (2015). Taking the Human Out of the Loop: A Review of Bayesian Optimization. Proceedings of the IEEE, 104(1):148–175.
Shi, F. G., Zhao, B., and Wang, S. Q. (1998). A New Theory for CMP with Soft Pads. In Proceedings of the IEEE 1998 International Interconnect Technology Conference (Cat. No. 98EX102), pages 73–75. IEEE.
Snoek, J., Larochelle, H., and Adams, R. P. (2012). Practical Bayesian Optimization of Machine Learning Algorithms. Advances in Neural Information Processing Systems, 25.
Wang, P., Gao, R. X., and Yan, R. (2017). A Deep Learning-Based Approach to Material Removal Rate Prediction in Polishing. CIRP Annals, 66(1):429–432.
Yeh, H. M. and Chen, K. S. (2010). Development of a Pad Conditioning Simulation Module with a Diamond Dresser for CMP Applications. The International Journal of Advanced Manufacturing Technology, 50(1):1–12.
Yu, H., Lin, C. C., Hsu, M. H., Chen, Y. T., Chen, K. W., Luoh, T., Yang, L. W., Yang, T., and Chen, K. C. (2021). CMP Process Optimization Engineering by Machine Learning.IEEE Transactions on Semiconductor Manufacturing, 34(3):280–285.
Yu, T., Li, Z., and Wu, D. (2019). Predictive Modeling of Material Removal Rate in Chemical Mechanical Planarization with Physics-Informed Machine Learning. Wear, 426:1430–1438.