Optical Proximity Correction Using Bidirectional Recurrent Neural Network With Attention Mechanism

Kwon, Yonghwi; Shin, Youngsoo

doi:10.1109/tsm.2021.3072668

Cited by 6 publications

(4 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To alleviate the long simulation runtime, numerous machine learning-based OPC models (MLOPC) have been proposed [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29]. Works from R. Frye [28] and P. Jedrasik [29] have implemented unsupervised neural networks for e-beam lithography and optical lithography for OPC, respectively.…”

Section: Related Workmentioning

confidence: 99%

“…S. Shim et al demonstrate the use of an artificial neural network for predicting the height of the resist after exposure [25], extending the efforts to etch proximity corrections (EPC) [26], using Fourier analysis in conjunction with MLOPC [11]. Y. Shin et al use recurrent neural networks for MLOPC [24]. P. Parashar et al use dimensionality reduction techniques to reduce network sizes [21].…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Optical Proximity Correction Using Machine Learning Assisted Human Decision

et al. 2023

View full text Add to dashboard Cite

Optical proximity correction (OPC) is a critical step in semiconductor manufacturing due to its high complexity and significant influence on the subsequent process steps. Conventional OPC using the Maxwell equation can become more and more challenging as a fully vectorized three-dimensional simulation is required for advanced technology nodes. Machine learning (ML) has been a promising alternative recently. This work proposes machine-learning-assisted human decision, which can be more in line with the clean room engineer's practice and can potentially surpass pure human decision and the pure machine learning approach. Using 10-step optimization in the photolithographic mask, the averaged mean square error (MSE) at the optimized cases are 6360 and 2101 for two-bar patterns and 7132 and 5931 for tri-line attackers when comparing pure human decision and ML-assisted human. The average MSEs at the first 3 steps are 26019 and 6023 for the two-bar pattern and 79979 and 7738 for the tri-line attacker when comparing pure ML and ML-assisted human. It is suggested that the strength of the ML-assisted human decision lies in the early-stage superiority over pure ML, flexibility, incorporating past experience, and a human sense that cannot be formulated concretely by statistical models.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Optical Proximity Correction Using Machine Learning Assisted Human Decision

et al. 2023

View full text Add to dashboard Cite

show abstract

“…ML and deep learning (DL) architectures and techniques have been applied in various tasks in the production line including overlay metrology, [3][4][5] wafer leveling and alignment, 6 defect detection and classification, 7,8 SEM images denoising, 9 and mask optimization, [10][11][12][13] and they have shown great improvement compared to conventional algorithms both in terms of accuracy and speed. In this research, we have demonstrated the application of our deep learning denoiser assisted framework towards enabling SEM contour extraction possible for all the edges in raw noisy DRAM SEM images with bit-line-periphery (BLP) and storage node landing pad (SNLP) patterns.…”

Section: Introductionmentioning

confidence: 99%

A deep learning facilitated approach for SEM image denoising towards improved contour extraction for 1D and 2D structures

Dey,

Wu,

Blanco

et al. 2023

38th European Mask and Lithography Conference (EMLC 2023)

View full text Add to dashboard Cite

“…Continuous attempts have led to some fruitful results in this area using deep convolution neural network (DCNN) architecture and bidirectional recurrent neural network 1 – 9 For machine learning OPC, the majority of the researches have been focused on designing some scheme that can measure the neighboring environment of a segment, then using the constructed feature vector from those measurements to predict the OPC amount of the segment through a trained neural network model 10 – 13 Such an approach may be adequate for hole layers, for which the types of rectangles (x dimension and y dimension) are very limited, consequently, the segmentation rules and control point setting rules are very limited.…”

Section: Introductionmentioning

confidence: 99%

Can optical proximity correction solution be learned? The learning limit and a general learning framework

Shi

Yan

et al. 2022

J. Micro/Nanopattern. Mats. Metro.

View full text Add to dashboard Cite

Background: Optical proximity correction (OPC) is an indispensable technology that has been propelling the advancement of computational lithography technology. To tightly control edge placement error (EPE) and maintain lithography process window, the demands on OPC computational resources and OPC turnaround time are growing rapidly with alarming acceleration. To tame the trend, machine learning technologies have been explored; however, an in-depth discussion on OPC solution learning limit is still lacking.Aim: We aim to present an in-depth discussion on OPC solution learning limit and propose a general machine learning OPC framework that can be extended to curvilinear mask OPC technology.Approach: In this study, we first investigate the machine learning OPC learning limit by examining noninverse lithography technology (non-ILT) OPC solution space characteristics inherited from edge segmentation and control point setting rules and then propose a general machine learning OPC framework that can take full advantage of deep convolution neural network (DCNN) learning capability while being able to preserve mask data high resolution.Results: With this machine learning OPC framework, we have achieved models with average absolute model error <1 nm for 14-nm metal layer. With single GPU, the average time for machine learning OPC models to produce results of 3840 nm × 3840 nm area is 8.74 ms for single channel input model and 12.65 ms for six channels input model. Conclusions:For non-ILT OPC solution, there is an intrinsic learning limit inherited from edge segmentation rules. Machine learning OPC models should be content with learning low order OPC solutions. This intrinsic learning limit of non-ILT OPC solution may diminish for ILT OPC solution when the constraint on degrees of freedom of OPC solution is lifted. The machine learning OPC framework we proposed is general and extendable to curvilinear OPC technology.

show abstract

Optical Proximity Correction Using Bidirectional Recurrent Neural Network With Attention Mechanism

Cited by 6 publications

References 11 publications

Optical Proximity Correction Using Machine Learning Assisted Human Decision

Optical Proximity Correction Using Machine Learning Assisted Human Decision

A deep learning facilitated approach for SEM image denoising towards improved contour extraction for 1D and 2D structures

Can optical proximity correction solution be learned? The learning limit and a general learning framework

Contact Info

Product

Resources

About