Topology-Driven Neural Defect Pattern Discovery in Silicon Wafers
DOI:
https://doi.org/10.5281/zenodo.20591504Keywords:
Semiconductor wafer inspection; defect pattern clustering; development and manufacturing of integrated circuits; supervised learning; convolutional neural networks; topological data analysis; decisionsupport; structural heat-damping composite materials; lab-scale manufacture; controlled heat-damping properties.Abstract
Semiconductor wafers undergo regular pattern inspections that are essential for assessing production quality. Recent years have seen the rise of deep-learning techniques for automating pattern recognition. They remain underexplored, however, for specific inspection tasks such as clustering similar defect patterns. TDA-based methods have been proposed that utilize TDA to delimit defect patterns and a Gaussian mixture model to perform clustering. A deep neural network is leveraged for the underlying representation learning. A recent dataset of defect pattern images is explored to enable better recognition of such inspections. Results indicate that the method outperforms several baseline methods in clustering accuracy across multiple scenes.
Semiconductors perform information acquisition, processing, transmission, and control, and are called the “brain” of electronics. They experience regular pattern inspections at each phase of the manufacturing process, with the inspection results serving as important references for process evaluation and control. Take wafer patterns as an example: if a pattern has serious defects, it cannot be used in subsequent manufacturing processes and must be discarded. Therefore, fast and effective inspection of wafer patterns is important for determining product quality. Recently, deep-learning techniques have been introduced to automate the process of inspection pattern recognition. These tools have been successfully applied in many fields, but they have not yet been fully applied to specific semiconductor inspection tasks such as clustering the same type of defects. TDA-based methods have been proposed that first employ TDA to delimit defect patterns and then a Gaussian mixture model to perform clustering on the defect patterns for a given annealing result.
References
[1] Adams, H., Tausz, A., & Vejdemo-Johansson, M. (2017). Persistence images: A stable vector representation of persistent homology. Journal of Machine Learning Research, 18(1), 218–252.
[2] Aksoy, S., & Haralick, R. M. (2001). Feature normalization and likelihood-based similarity measures for image retrieval. Pattern Recognition Letters, 22(5), 563–582.
[3] Attali, D., Glisse, M., Hornus, S., & Lazarus, F. (2013). Persistence-sensitive simplification of functions on surfaces in linear time. In Proceedings of the 29th Annual Symposium on Computational Geometry (pp. 405–414). ACM.
[4] Bae, J., Kim, H., Lee, J., & Park, S. (2020). Deep learning-based defect classification for semiconductor manufacturing: A comprehensive survey. IEEE Access, 8, 171219–171236.
[5] Bendich, P., Marron, J. S., Miller, E., Pieloch, A., & Skwerer, S. (2016). Persistent homology analysis of brain artery trees. Annals of Applied Statistics, 10(1), 198–218.
[6] Bubenik, P. (2015). Statistical topological data analysis using persistence landscapes. Journal of Machine Learning Research, 16, 77–102.
[7] Carlsson, G. (2009). Topology and data. Bulletin of the American Mathematical Society, 46(2), 255–308.
[8] Carlsson, G., Ishkhanov, T., de Silva, V., & Zomorodian, A. (2008). On the local behavior of spaces of natural images. International Journal of Computer Vision, 76(1), 1–12.
[9] Chazal, F., de Silva, V., Glisse, M., & Oudot, S. (2016). The structure and stability of persistence modules. SpringerBriefs in Mathematics. Springer.
[10] Chen, C., Ni, X., Bai, Y., Wang, Y., & Wang, Y. (2015). A practical algorithm for computing persistent homology of cubical data. In Proceedings of the 2015 IEEE International Conference on Data Mining Workshops (pp. 127–134). IEEE.
[10] Chollet, F. (2017). Xception: Deep learning with depthwise separable convolutions. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 1251–1258).
[11] Chung, Y.-M., Kim, J., & Lee, S. (2020). Wafer defect pattern recognition and classification using deep convolutional neural networks. IEEE Transactions on Semiconductor Manufacturing, 33(3), 430–437.
[12] Cireşan, D. C., Meier, U., Masci, J., Gambardella, L. M., & Schmidhuber, J. (2012). Flexible, high performance convolutional neural networks for image classification. In Proceedings of the 22nd International Joint Conference on Artificial Intelligence (pp. 1237–1242). AAAI Press.
[13] Cohen-Steiner, D., Edelsbrunner, H., & Harer, J. (2007). Stability of persistence diagrams. Discrete & Computational Geometry, 37(1), 103–120.
[14] Costan, A., & Courtine, P. (2021). Automated defect inspection in semiconductor manufacturing using deep learning: A review. Computers in Industry, 123, 103306.
[15] Cui, Z., Chen, W., & Chen, Y. (2016). Multi-scale convolutional neural networks for time series classification. arXiv (preprint).
(Note: arXiv is published as a public preprint archive; include only if acceptable in your institution.)
[16] Dau, H. A., Keogh, E., Kamgar, K., Yeh, C.-C. M., Zhu, Y., Gharghabi, S., Ratanamahatana, C. A., & Yanping. (2019). The UCR time series archive. IEEE/CAA Journal of Automatica Sinica, 6(6), 1293–1305.
[17] De Silva, V., & Ghrist, R. (2007). Coverage in sensor networks via persistent homology. Algebraic & Geometric Topology, 7(1), 339–358.
[18] Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009). ImageNet: A large-scale hierarchical image database. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 248–255). IEEE.
[19] Edelsbrunner, H., & Harer, J. (2010). Computational topology: An introduction. American Mathematical Society.
[20] Edelsbrunner, H., Letscher, D., & Zomorodian, A. (2002). Topological persistence and simplification. Discrete & Computational Geometry, 28(4), 511–533.
[21] Frosini, P., & Landi, C. (1999). Size theory as a topological tool for computer vision. Pattern Recognition and Image Analysis, 9(4), 596–603.
[22] Ghahramani, Z. (2015). Probabilistic machine learning and artificial intelligence. Nature, 521(7553), 452–459.
[23] Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep learning. MIT Press.
[24] He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 770–778).
[25] Hofer, C., Kwitt, R., Niethammer, M., & Uhl, A. (2017). Deep learning with topological signatures. In Advances in Neural Information Processing Systems (pp. 1634–1644).
[26] Ismail Fawaz, H., Forestier, G., Weber, J., Idoumghar, L., & Muller, P.-A. (2019). Deep learning for time series classification: A review. Data Mining and Knowledge Discovery, 33(4), 917–963.
[27] Jung, H., Lee, S., Kim, J., & Park, K. (2021). Unsupervised wafer map defect pattern clustering using deep feature embeddings. IEEE Transactions on Semiconductor Manufacturing, 34(4), 512–521.
[28] Kingma, D. P., & Welling, M. (2014). Auto-encoding variational Bayes. In Proceedings of the 2nd International Conference on Learning Representations.
[29] Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. In Advances in Neural Information Processing Systems (pp. 1097–1105).
[30] Lai, Y.-H., Tsai, C.-Y., & Chen, S.-L. (2019). A deep learning approach for wafer map defect pattern classification. Journal of Intelligent Manufacturing, 30(6), 2325–2337.
[31] LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444.
[32] Li, J., Qiao, Y., Wang, Y., & Zhang, H. (2020). A review of defect detection and analysis in semiconductor manufacturing using machine learning. IEEE Transactions on Industrial Informatics, 16(10), 6431–6443.
[33] Long, J., Shelhamer, E., & Darrell, T. (2015). Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 3431–3440).
[34] Marian, Z., Somnath, S., Kalinin, S. V., & Vasudevan, R. K. (2022). Machine learning for defect detection and analysis in manufacturing. Nature Reviews Materials, 7, 563–578.
[35] Moor, M., Horn, M., Rieck, B., & Borgwardt, K. (2020). Topological Autoencoders. In Proceedings of the 37th International Conference on Machine Learning (pp. 7045–7054). PMLR.
[36] Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., Vanderplas, J., Passos, A., Cournapeau, D., Brucher, M., Perrot, M., & Duchesnay, E. (2011). Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12, 2825–2830.
[37] Saul, N., & Tralie, C. (2019). Scalable visualization of persistent homology for high-dimensional data. In Proceedings of the 27th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (pp. 447–450). ACM.
[38] Somnath, S., & Kalinin, S. V. (2021). Automated discovery of structure–property relationships using deep learning in microscopy and manufacturing. Advanced Intelligent Systems, 3(9), 2000275.
[39] Zomorodian, A., & Carlsson, G. (2005). Computing persistent homology. Discrete & Computational Geometry, 33(2), 249–274.
Additional Files
Published
Data Availability Statement
The dataset used in this study was provided by a Taiwan-based semiconductor manufacturing company and is not publicly available due to proprietary restrictions. A public wafer defect dataset (SE-WWI) was also used for comparative experiments.
