Abstract
Manufacturing of integrated circuits involves many sequential processes executed to nanoscale tolerances, and the yield depends on the often-unmeasured quality of intermediate steps. Taking physical quality measurements in this high-throughput industry can be expensive and time-consuming. Instead, we seek to predict product quality characteristics using readily available sensor readings of the tool environment during processing of each wafer and employ Gaussian Process Regression (GPR) paradigm to realize this Virtual Metrology (VM) concept. Convergence of the GPR based VM estimation of product quality is hastened through an active sampling scheme, whereby the predictive uncertainty of the GPR model informs which wafer’s quality to measure next in order to obtain maximal additional information for the VM model. We evaluate these methods using a large dataset collected from a plasma-enhanced chemical vapor deposition (PECVD) process, with relevant tool sensor readings and the corresponding physical measurements of mean film thicknesses for 32,000 wafers. By selecting which wafers to physically measure for VM model updates, the GPR based VM method achieves ~10% greater accuracy on average than the partial least squares based method.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
In deriving a GP, the underlying parametric model’s weights are integrated out, so the tunable parameters of the GP prior are referred to as hyperparameters to emphasize the non-parametric nature of this approach.
- 2.
Since the dependence of the latent function, f on the parameters, θ tends to be well-concentrated, approximating this dependence as a line through the MLE, \( \widehat{\theta } \) seems reasonable.
- 3.
The invariance of mutual information to homeomorphic transformations make this learning process robust to scaling or rotating of the current embedding, or to centering and scaling of the inputs and outputs.
- 4.
RMSE more heavily penalizes large predictive errors than mean absolute error (MAE), and typically yields clearer outliers.
References
Ringwood, J.V., Lynn, S., Bacelli, G., Ma, B., Ragnioli, E., McLoon, S.: Estimation and control in semiconductor etch: practice and possibilities. IEEE Trans. Semicond. Manuf. 23(1), 87–98 (2010)
An-Jhih, S., Jeng, J.-C., Huang, H.-P., Cheng-Ching, Yu., Hung, S.-Y., Chao, C.-K.: Control relevant issues in semiconductor manufacturing: overview with some new results. Control Eng. Pract. 15(10), 1268–1279 (2007)
Yang, Y., Wang, M., Kushner, M.J.: Progress, opportunities and challenges in modeling of plasma etching. In: Proceedings of the 2008 IEEE International Interconnect Technology Conference (IITC), Burlingame, CA, USA, 1–4 June 2008 (2008). https://doi.org/10.1109/iitc.2008.4546934
Hirai, T., Kano, M.: Adaptive virtual metrology design for semiconductor dry etching process through locally weighted partial least squares. IEEE Trans. Semicond. Manuf. 28(2), 137–144 (2015)
Bleakie, A., Djurdjanovic, D.: Growing structure multiple model system for quality estimation in manufacturing processes. IEEE Trans. Semicond. Manuf. 29(2), 79–97 (2016)
Lynn, S.A., Ringwood, J., MacGearailt, N.: Global and local virtual metrology models for a plasma etch process. IEEE Trans. Semicond. Manuf. 25(1), 94–103 (2012)
Lee, S., Kang, P., Cho, S.: Probabilistic local reconstruction for k-NN regression and its application to virtual metrology in semiconductor manufacturing. Neurocomputing 131(5), 427–439 (2014)
Rasmussen, C.E.: Gaussian Processes for Machine Learning. MIT Press, Cambridge (2006). ISBN 026218253X
Garnett, R., Osborne, M.A., Hennig, P.: Active learning of linear embeddings for gaussian processes. In: Proceedings of the 13th Conference on Uncertainty in Artificial Intelligence, Quebec City, Quebec, Canada, 23–27 July 2014, pp. 230–239 (2014)
Osborne, M.A., Garnett, R., Roberts, S.J.: Gaussian processes for global optimization. In: Proceedings of the 3rd International Conference on Learning and Intelligent Optimization, 14–18 January 2009, Trento, Italy (2009)
Neal, R.M.: Bayesian Learning for Neural Networks. Springer Science and Business Media, New York (1996)
Wold, S., Sjöström, M., Eriksson, L.: PLS-regression: a basic tool of chemometrics. Chemometr. Intell. Lab. Syst. 58(2), 109–130 (2001)
Kullback, S., Leibler, R.A.: On information and sufficiency. Ann. Math. Stat. 22(1), 79–86 (1951)
McInerney, E.J.: Chemical vapor deposition. In: Geng, H. (ed.) Semiconductor Manufacturing Handbook. McGraw-Hill Inc., New York (2005). Chapter 14
Bleakie, A., Djurdjanovic, D.: Feature extraction, condition monitoring and fault modeling in semiconductor manufacturing systems. Comput. Ind. 64(3), 203–213 (2013)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer International Publishing AG, part of Springer Nature
About this paper
Cite this paper
Darwin, T., Garnett, R., Djurdjanovic, D. (2018). Gaussian Process Regression for Virtual Metrology of Microchip Quality and the Resulting Selective Sampling Scheme. In: Ni, J., Majstorovic, V., Djurdjanovic, D. (eds) Proceedings of 3rd International Conference on the Industry 4.0 Model for Advanced Manufacturing. AMP 2018. Lecture Notes in Mechanical Engineering. Springer, Cham. https://doi.org/10.1007/978-3-319-89563-5_19
Download citation
DOI: https://doi.org/10.1007/978-3-319-89563-5_19
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-89562-8
Online ISBN: 978-3-319-89563-5
eBook Packages: EngineeringEngineering (R0)