ML Photolithography Mask Simulation for 3D Effects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for simulating aerial images of photolithography masks are either time-consuming and inaccurate or require extensive computational resources, failing to effectively account for mask 3D effects and short wavelengths.
Innovation Solution
A computer-implemented method using a machine learning model to simulate the propagation of electromagnetic waves through a detailed model of a photolithography mask, incorporating absorber and mask carrier sections, and applying a simulation of the imaging process to generate an aerial image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If rigorous simulation methods (FDTD, RCWA) are used to simulate aerial images, then accuracy is improved, but computation time and computational resources increase significantly
Solution Approach 1:
The patent creates a machine learning model that learns from rigorous simulation data and produces aerial image simulations that copy the accuracy of FDTD/RCWA methods but with much faster computation. The ML model is trained on training data generated by rigorous simulation methods, allowing it to replicate their accuracy without requiring the same computational intensity during actual use.
Solution Approach 2:
The patent changes the fundamental approach from direct physical simulation to data-driven prediction. By transforming the simulation problem into a pattern recognition task where the ML model maps mask models to aerial images through learned parameters, it achieves both speed and accuracy without following the computationally intensive physical simulation steps.
2Loss of time
If thin element approximation is used for fast aerial image simulation, then computation time is reduced, but accuracy deteriorates due to inability to account for mask 3D effects
Solution Approach 1:
The patent changes the input parameters to include comprehensive 3D mask geometry information (absorber section height, mask carrier section thickness, multilayer structure) rather than relying on thin element approximations. This allows the ML model to capture mask 3D effects while maintaining fast computation speeds.
Solution Approach 2:
The ML model copies the physical accuracy of rigorous simulation methods by being trained on their output data, while avoiding their computational burden. It learns the complex relationships between 3D mask structures and aerial images without performing the intensive physical calculations required by FDTD or RCWA.
3Measurement precision
If detailed 3D mask models are used to account for mask 3D effects, then accuracy is improved, but device complexity and computational requirements increase
Solution Approach 1:
The patent uses a machine learning model that copies the results of complex rigorous simulations without replicating their complexity. The ML model takes detailed 3D mask models as input but processes them through simplified pattern recognition operations, reducing the computational complexity required to achieve the same accuracy.
Solution Approach 2:
The patent replaces the mechanical/physical simulation process (solving Maxwell's equations numerically) with an information-processing approach (machine learning pattern recognition). This substitution dramatically reduces computational complexity while maintaining accuracy through data-driven learning of the physical relationships.
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
The method achieves accurate and fast simulation of aerial images, reducing computation time and memory requirements while effectively modeling mask 3D effects and short wavelengths, thereby improving photolithography mask design and defect detection.
Implementation Method 1
simulating the propagation of the incident electromagnetic waves through the model of the photolithography mask using a machine learning model, wherein the machine learning model maps the model of the photolithography mask to a representation of an electromagnetic field generated by the incident electromagnetic waves
Implementation Method 2
the grating comprising absorber structures and non-absorber structures forming a pattern on at least a portion of the mask carrier
Data Source
AI summary
The invention relates to a computer implemented method for simulating an aerial image of a model of a photolithography mask illuminated by incident electromagnetic waves, the method comprising: obtaining the model of the photolithography mask; simulating the propagation of the incident electromagnetic waves through the model of the photolithography mask using a machine learning model, wherein the machine learning model maps the model of the photolithography mask to a representation of an electromagnetic field generated by the incident electromagnetic waves on the photolithography mask; obtaining the aerial image of the model of the photolithography mask by applying a simulation of an imaging process. The invention also relates to corresponding computer programs, computer-readable media and systems.