Lithography Model Calibration for Proximity Effect Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lithography techniques face challenges in accurately calibrating models to compensate for the proximity effect, especially at extreme low k1 values, leading to significant geometric differences between target and printed patterns, which affects critical dimensions and topology.
Innovation Solution
A method for calibrating a lithography model by iteratively adjusting parameters to minimize a demerit function, which includes calculating differences between simulated and measured image intensities and critical dimensions, with optional normalization and weighting, and excluding non-printing features from the calculation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If optical proximity correction and resolution enhancement techniques are applied, then the fidelity of printed patterns is improved, but significant geometric differences remain between target layout and printed pattern at extreme low k1 values
Solution Approach 1:
The patent applies parameter changes by iteratively adjusting lithography model parameters (such as proximity effect coefficients, focus exposure matrix parameters, and resist model parameters) to minimize the difference between simulated and measured patterns. This allows the model to adapt to extreme low k1 conditions and capture the proximity effect more accurately, reducing geometric differences without requiring traditional OPC/RET techniques alone.
Solution Approach 2:
The patent implements feedback by using measured critical dimensions and pattern data from actual lithography processes to calibrate and refine the lithography model. The model parameters are continuously adjusted based on comparison between simulated and measured results, creating a closed-loop system that improves accuracy at extreme k1 values where traditional techniques fail.
2Manufacturing precision
If the k1 factor is decreased below 0.35, then the resolution and density of printed patterns is improved, but the proximity effect becomes strong and maintaining sufficient yield becomes increasingly difficult
Solution Approach 1:
The patent replaces traditional geometric-based OPC/RET approaches with a physics-based lithography model that explicitly simulates the proximity effect through parameterized equations. This substitution allows direct modeling of the physical processes (electron scattering, light diffusion, molecule diffusion) that cause proximity effects, enabling accurate prediction and control at extreme low k1 values where geometric corrections become insufficient.
Solution Approach 2:
The patent changes the approach from geometric corrections to physical parameter modeling by adjusting proximity effect coefficients, focus-exposure matrix parameters, and resist development parameters. This parameter-based approach directly addresses the physical mechanisms of proximity effects, allowing reliable process control at k1 < 0.35 where traditional methods break down.
3Device complexity
If traditional geometric considerations are used for OPC, then the calibration process is simpler, but the accuracy of critical dimension prediction deteriorates at extreme low k1 values
Solution Approach 1:
The patent transitions from simple geometric parameters to a comprehensive set of physical parameters including proximity effect coefficients, focus-exposure matrix elements, and resist model parameters. This parameter expansion enables accurate critical dimension prediction at extreme low k1 values by capturing the underlying physics of the lithography process, despite increased calibration complexity.
Solution Approach 2:
The patent introduces feedback mechanisms that compare simulated critical dimensions against measured values and use this information to iteratively refine model parameters. This feedback loop compensates for the increased complexity by automatically adjusting parameters to achieve accurate predictions, making the complex physical model as usable as simple geometric methods.
Data Source
AI summary
The present invention provides a method for calibrating a computational model of a lithography process by calculating a demerit function using an intensity measurement at a location of a wafer; and calibrating the lithography model or a mask making model by determining values of parameters of the computational model using the calculated demerit function. The method may also use a second demerit function that is defined by the sum of squares of differences between a simulated and measured critical dimensions of a feature on the wafer.