Optical Proximity Correction Using Process Window Error Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical proximity correction (OPC) techniques in semiconductor lithography optimize feature printing only at best exposure and focus conditions, failing to account for variations across the full process window, leading to defects at other parameter values within the desired process range.
Innovation Solution
The implementation of an OPC strategy that uses process window information to optimize reticle design by adjusting error functions across the full range of lithographic process parameters, including focus, exposure, and other critical factors, to ensure high-yield feature printing across the entire process window.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If OPC optimizes feature printing only at best exposure and focus conditions, then printing accuracy at optimal parameters is improved, but printing robustness across the full process window deteriorates
Solution Approach 1:
The patent applies parameter changes by modifying the error function to incorporate multiple process parameters (exposure dose, focus, partial coherence factor, numerical aperture) rather than optimizing for a single best condition. This allows the OPC to adjust reticle features to compensate for variations across the entire process window, transforming the optimization from a single-point solution to a multi-parameter robust solution.
Solution Approach 2:
The patent implements preliminary action by pre-calculating and storing process window information (PWI) that characterizes how printing errors vary across different exposure and focus conditions. This pre-computed data is then used during OPC to guide the optimization process, allowing the system to anticipate and compensate for process variations before actual lithography occurs.
2Reliability
If OPC uses process window information to optimize across full parameter range, then printing robustness is improved, but computational complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the continuous process window into discrete sampling points across the parameter space (exposure dose, focus, partial coherence factor, numerical aperture). The error function is constructed by summing contributions from each sampled condition, transforming a complex continuous optimization problem into a manageable discrete summation that can be computed efficiently.
Solution Approach 2:
The patent uses copying by creating a simplified computational model of the lithography process that replicates the essential physics without requiring full-wave electromagnetic simulations. The error function uses approximate models (e.g., coherent imaging models with scalar diffraction) that capture the dominant effects while being computationally tractable, allowing rapid optimization across the process window.
Data Source
AI summary
Optical proximity correction methods and apparatus are disclosed. A simulated geometry representing one or more printed features from a reticle is generated using an optical proximity correction (OPC) model that takes into account a reticle design and one or more parameters from a process window of a stepper. An error function is formed that measures a deviation between the simulated geometry and a desired design of the one or more printed features. The error function takes into account parameters (p0 . . . pJ) from across the process window in addition to, or in lieu of, a best focus and a best exposure for the stepper. The reticle design is adjusted in a way that reduces the deviation as measured by the error function, thereby producing an adjusted reticle design.