Multi-Beam Lithography Neural Network Shot Conversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current optical lithography techniques face challenges in accurately transferring patterns with features smaller than the light wavelength used, requiring complex sub-resolution assist features and increasing mask complexity, which complicates the manufacturing process and increases costs.
Innovation Solution
A method using multi-beam charged particle beam lithography, where a neural network is employed to convert single-beam shot data into multi-beam shot data, optimizing the pattern formation on a surface by generating a set of multi-beam shots that match a calculated image within a predetermined tolerance, utilizing computing hardware processors for efficient data preparation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If optical lithography is used to transfer patterns with features smaller than the light wavelength, then pattern transfer capability is improved, but mask complexity increases due to required sub-resolution assist features
Solution Approach 1:
The patent replaces optical lithography with charged particle beam lithography (electron beam or ion beam). This substitution eliminates the diffraction limit constraint of optical methods, enabling direct writing of sub-wavelength patterns without requiring complex sub-resolution assist features or advanced optical proximity correction, thus resolving the contradiction between pattern transfer capability and mask complexity
Solution Approach 2:
The patent employs multi-beam charged particle lithography where a single electron or ion beam is divided into multiple beamlets through an aperture array. This segmentation allows simultaneous exposure of multiple pattern regions, improving throughput while maintaining the ability to write complex sub-wavelength patterns directly without complex masks
2Ease of manufacture
If conventional optical lithography with 193 nm wavelength is used, then manufacturing process is simpler, but pattern resolution is limited by the light wavelength
Solution Approach 1:
The patent changes the fundamental parameter of the lithography method from optical (193 nm wavelength) to charged particle beam (electron or ion beam). This parameter change eliminates the diffraction limit that constrains optical lithography resolution, enabling direct writing of patterns at resolutions below the optical wavelength while using a relatively simple direct-write approach without complex optical correction systems
3Device complexity
If single-beam charged particle lithography is used, then equipment complexity is lower, but writing speed and productivity are insufficient for high-volume manufacturing
Solution Approach 1:
The patent divides a single electron or ion beam into multiple beamlets using an aperture array, creating a multi-beam system. This allows simultaneous exposure of multiple pattern regions, increasing writing speed by a factor proportional to the number of beamlets while adding only moderate complexity through the aperture array and beam control electronics
Solution Approach 2:
The patent combines multiple beam exposure operations into a single simultaneous exposure event. By merging the functionality of multiple single-beam exposures into one multi-beam operation, the system achieves high throughput suitable for high-volume manufacturing while maintaining relatively simple equipment architecture
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the accuracy and efficiency of pattern transfer, reduces the need for complex OPC features, and decreases the overall complexity of the mask data preparation, thereby improving the precision and reducing the costs associated with advanced lithography processes.
Implementation Method 1
a charged particle beam is used to expose a resist-coated surface
Data Source
AI summary
Methods for fracturing or mask data preparation are disclosed in which a set of single-beam charged particle beam shots is input; a calculated image is calculated using a neural network, from the set of single-beam charged particle beam shots; and a set of multi-beam shots is generated based on the calculated image, to convert the set of single-beam charged particle beam shots to the set of multi-beam shots which will produce a surface image on the surface. Methods for training a neural network include inputting a set of single-beam charged particle beam shots; calculating a set of calculated images using the set of single-beam charged particle beam shots; and training the neural network with the set of calculated images.