Deep learning denoising of SEM images towards noise-reduced LER measurements

Abstract As chip sizes decrease and node dimensions break the sub-10 nm barrier, Line Edge Roughness (LER) metrology becomes a critical issue for the semiconductor research and industry. Scanning Electron Microscopy (SEM) imaging being the widely used tool for LER metrology suffers from the presence of noise that degrades measurement accuracy. To solve this issue without damaging the measured pattern, the applicability of deep Convolutional Neural Networks (CNNs) is explored, tackling the problem at the image level. The SEM image Denoising model (SEMD) is trained on synthesized image data to detect the variations of noise and provides as an output a denoised image of the pattern. The results are presented and compared with the state-of-art predictions, showing the effectiveness and enhanced performance of the SEMD method.

[1]  Atsuko Yamaguchi,et al.  Bias-free measurement of LER/LWR with low damage by CD-SEM , 2006, SPIE Advanced Lithography.

[2]  Soo-Young Lee,et al.  Noise filtering for accurate measurement of line edge roughness and critical dimension from SEM images , 2016 .

[3]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[4]  Vassilios Constantoudis,et al.  Line edge roughness measurement through SEM images: effects of image digitization and their mitigation , 2017, European Mask and Lithography Conference.

[5]  P Cizmar,et al.  Simulated SEM images for resolution measurement. , 2008, Scanning.

[6]  Lei Zhang,et al.  Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising , 2016, IEEE Transactions on Image Processing.

[7]  David Z. Pan,et al.  Electrical impact of line-edge roughness on sub-45-nm node standard cells , 2010 .

[8]  Geoffrey E. Hinton,et al.  Rectified Linear Units Improve Restricted Boltzmann Machines , 2010, ICML.

[9]  Evangelos Gogolides,et al.  Sidewall roughness in nanolithography: origins, metrology and device effects , 2014 .

[10]  Jong Chul Ye,et al.  A deep convolutional neural network using directional wavelets for low‐dose X‐ray CT reconstruction , 2016, Medical physics.

[11]  C. Mack,et al.  Unbiased roughness measurements: Subtracting out SEM effects , 2018 .

[12]  Wangmeng Zuo,et al.  Learning Deep CNN Denoiser Prior for Image Restoration , 2017, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[13]  Angeliki Tserepi,et al.  Quantification of line-edge roughness of photoresists. I. A comparison between off-line and on-line analysis of top-down scanning electron microscopy images , 2003 .

[14]  Lei Zhang,et al.  Nonlocally Centralized Sparse Representation for Image Restoration , 2013, IEEE Transactions on Image Processing.

[15]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[16]  Makoto Suzuki,et al.  SEM image prediction based on modeling of electron-solid interaction , 2017, Advanced Lithography.

[17]  J. S. Villarrubia,et al.  Unbiased estimation of linewidth roughness , 2005, SPIE Advanced Lithography.

[18]  Ohad Shamir,et al.  Better Mini-Batch Algorithms via Accelerated Gradient Methods , 2011, NIPS.

[19]  Evangelos Gogolides,et al.  Noise-free estimation of spatial line edge/width roughness parameters , 2009, Advanced Lithography.

[20]  Claus Nebauer,et al.  Evaluation of convolutional neural networks for visual recognition , 1998, IEEE Trans. Neural Networks.

[21]  Evangelos Gogolides,et al.  Line edge roughness metrology: recent challenges and advances toward more complete and accurate measurements , 2018, Journal of Micro/Nanolithography, MEMS, and MOEMS.