Proximity Effect Correction via Probabilistic Deconvolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current electron lithography techniques face challenges in accurately modeling and correcting proximity effects at critical dimensions below 22 nm due to the limitations of existing dose modulation and geometric correction methods, particularly for complex patterns, leading to sub-optimal design and exposure times.
Innovation Solution
A method combining a specific point spread function (PSF) with an iterative probabilistic deconvolution process using the Lucy-Richardson method to model forward scattering effects, allowing for joint optimization of radiated dose and geometry parameters, and enabling exposure at multiple levels to correct proximity effects automatically.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If separate calculation of dose modulation and geometric correction parameters is used, then design time and exposure time are reduced, but manufacturing precision deteriorates for complex patterns
Solution Approach 1:
The patent combines separate dose modulation and geometric correction calculations into a unified proximity effect correction process. By integrating these previously independent operations, the system achieves both time efficiency and high manufacturing precision for complex patterns through simultaneous optimization of both parameters.
Solution Approach 2:
The invention transforms the correction approach by changing from separate parameter calculations to a unified multi-parameter optimization. The system simultaneously adjusts dose modulation parameters and geometric correction parameters based on pattern complexity, enabling adaptive parameter selection that maintains precision while managing computational efficiency.
2Measurement precision
If inverse Fourier transforms with classic two-Gaussian PSF are used, then deconvolution is achieved, but manufacturing precision deteriorates at critical dimensions of one to two tens of nanometers due to high spatial frequency cutoff
Solution Approach 1:
The patent applies different PSF characteristics to different spatial frequency ranges. The improved PSF model accurately represents forward scattering effects at high spatial frequencies (critical dimensions of 1-20 nm) while maintaining appropriate modeling of backscattering at lower frequencies, enabling precise deconvolution across the entire frequency spectrum without the cutoff limitation of classic approaches.
Solution Approach 2:
The invention uses a composite PSF model that combines multiple Gaussian functions with different parameters to represent different scattering regimes. This composite approach integrates both forward scattering (dominant at short distances) and backscattering (dominant at longer distances) characteristics, providing accurate deconvolution for complex patterns at nanometer-scale critical dimensions.
3Manufacturing precision
If classic two-Gaussian PSF is used, then both forward and backward scattering are modeled, but device complexity increases and manufacturing precision deteriorates for forward scattering dominance at short distances
Solution Approach 1:
The patent segments the scattering modeling into distinct components: forward scattering (modeled by one or more Gaussians with specific parameters) and backscattering (modeled by separate Gaussian functions). This segmentation allows the system to apply appropriate modeling complexity only where needed, reducing overall model complexity while maintaining manufacturing precision for forward scattering-dominant scenarios at short distances.
Solution Approach 2:
The invention implements a dynamic PSF model where the contribution of forward and backward scattering components is adjusted based on the specific pattern geometry and critical dimensions. For patterns with small critical dimensions where forward scattering dominates, the model dynamically emphasizes the forward scattering component, reducing unnecessary computational complexity while maintaining accuracy.
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 enables the production of faithful patterns with optimal radiated images, improving resolution and writing speed by automatically generating correction rules for complex patterns with smaller critical dimensions, overcoming the limitations of previous methods.
Implementation Method 1
the interaction of the electron beam with the target is notably affected by a diffusion of the electrons around the initial trajectory (effect of forward scattering or forward scattering)
Implementation Method 2
as well as by backscattering (back scattering effect)
Data Source
AI summary
The invention pertains to a lithography method by radiation having critical dimensions of around ten nanometers. The method of the invention makes it possible to carry out the correction of the effects of proximity by the joint optimization of the modulation of loads and of the geometric corrections. For this purpose, a deconvolution of the pattern to be engraved is carried out using an iterative method modeling the interactions of the radiation with the resin substrate by means of a joint probability distribution. Advantageously, when the tool for exposing the substrate has a formed beam, the pattern to be engraved is cut with contrasting levels, and then the deconvoluted image is vectorized and broken down before carrying out the exposure step. In an advantageous embodiment, the method of the invention is applied to at least two character cells that are exposed in a multi-pass cell projection method.