Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A Deep Learning Approach to Computational Lithography

A deep learning and computational lithography technology, applied in optics, microlithography exposure equipment, optomechanical equipment, etc., can solve problems such as limiting the scope of application, avoid training errors or training failures, simplify the training process, and improve efficiency. Effect

Active Publication Date: 2020-08-21
BEIJING INSTITUTE OF TECHNOLOGYGY
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, most practical lithography systems use partially coherent illumination, thus limiting the scope of application of the above techniques

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A Deep Learning Approach to Computational Lithography
  • A Deep Learning Approach to Computational Lithography
  • A Deep Learning Approach to Computational Lithography

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0032] The present invention will be described in detail below with reference to the accompanying drawings and examples.

[0033] Please refer to figure 1 , figure 1 A deep learning method for computational lithography provided by an embodiment of the present invention is shown, and its basic scheme is:

[0034] S1. Based on the lithographic imaging model, construct the mask graphic update formula in the gradient iterative algorithm; expand the iterative process of the gradient iterative algorithm, intercept the first K steps of iterations, and use each step of the iterative process as a layer in MCNN to create a K layer feed-forward convolutional neural network. The present invention refers to this model-based convolutional neural network as a model-driven convolutional neural network MCNN. The feed-forward MCNN can be considered as an encoder.

[0035] The input of the encoder is a given ideal circuit layout, and its output is the corresponding OPC mask pattern. Here, "...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a deep learning method for computational lithography, which can improve the optimization speed and convergence performance of the traditional OPC method, and can be used for both deep ultraviolet DUV lithography and extreme ultraviolet EUV lithography. Firstly, the gradient iterative algorithm is expanded and its first steps are intercepted, combined with the imaging model of the lithography system and the photoresist model, a forward model-driven convolutional neural network MCNN, namely the encoder, is created. Create a decoder corresponding to the encoder, connect the output of the MCNN to the input of the decoder, use a certain distance between the input of the MCNN and the output of the decoder as a cost function, and train the parameters of the MCNN. The training set samples are input into MCNN, and the parameters in MCNN are optimized to minimize the error between the input of MCNN and the output of the decoder. After the training is completed, the MCNN is separated from the decoder, and the OPC mask graphics corresponding to other circuit layouts can be obtained by using the MCNN.

Description

technical field [0001] The present invention relates to the technical fields of computational imaging, computational lithography and deep learning, in particular to a computational lithography method based on a Model-driven Convolution Neural Network (MCNN for short). Background technique [0002] Photolithography is one of the core technologies used to manufacture VLSI. The modern integrated circuit manufacturing industry continues to develop basically in accordance with Moore's Law. The chip's critical dimension (CD) is constantly shrinking. The development process from DUV (referred to as DUV) lithography to extreme ultraviolet (Extreme Ultraviolet, EUV) lithography. In different lithography technology nodes, different lithography systems may be required for integrated circuit fabrication. [0003] Projection lithography is widely used in the current VLSI mass production process due to its high resolution, no contamination of the mask, and good repeatability. In the DU...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G03F7/20G06N3/04
CPCG03F7/705G03F7/70508G06N3/045
Inventor 马旭郑现强
Owner BEIJING INSTITUTE OF TECHNOLOGYGY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com