Optical Proximity Correction Using Weighted Average Error Retargeting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current optical proximity correction (OPC) techniques struggle to ensure process window tolerance and fidelity in deep sub-wavelength lithography, where feature sizes approach or exceed the wavelength of light used, leading to non-linear pattern transfer and complex interactions between features, resulting in reduced process window and increased errors across varying focus and exposure conditions.
Innovation Solution
The method involves applying optical proximity corrections using a model of the lithography process at nominal conditions to produce a first post-OPC layout, simulating the process across multiple conditions to determine weighted average errors, and applying retarget values to adjust edge segments to optimize critical dimensions and edge placement errors, thereby expanding the process window and improving pattern fidelity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If optical proximity correction is applied using a single nominal process model, then the correction can be calculated efficiently, but the process window depth of focus and tolerance is reduced
Solution Approach 1:
The patent segments the process window into multiple discrete process conditions (e.g., multiple focus offsets and exposure doses). Instead of using a single nominal model, the system performs OPC calculations for each segmented process condition, thereby capturing the full range of process variations and expanding the effective process window while maintaining computational efficiency through automated batch processing.
Solution Approach 2:
The patent changes the process model parameters dynamically by selecting different process conditions (focus offset, exposure dose) from a predefined set. The system iterates through multiple parameter combinations, performing OPC corrections for each condition and combining the results to achieve a robust correction that works across the entire process window, not just at nominal conditions.
2Manufacturing precision
If optical proximity correction is optimized for a specific process condition, then the correction accuracy at that condition is maximized, but the correction performance degrades under varying process conditions
Solution Approach 1:
The patent performs preliminary OPC corrections for multiple hypothetical process conditions before actual lithography. By pre-calculating corrections for various focus offsets and exposure doses, the system prepares a comprehensive correction database that can be applied to actual process variations, ensuring accurate critical dimensions across all conditions rather than just at the nominal point.
Solution Approach 2:
The system incorporates feedback by comparing simulated resist images at different process conditions against the desired pattern. The OPC corrections are iteratively adjusted based on this feedback, and the results are validated through simulation to ensure that the corrections work across the entire process window, not just at a single condition.
3Reliability
If multiple process conditions are simulated to expand process window, then the process window depth of focus is increased, but the computational time and complexity increase
Solution Approach 1:
The patent performs all OPC corrections and simulations in advance during the mask design phase. By pre-calculating corrections for multiple process conditions and storing them in a lookup table or database, the system eliminates the need for time-consuming iterative calculations during actual lithography processing, thereby expanding the process window without increasing real-time computational time.
Solution Approach 2:
The system creates virtual copies of the resist image at different process conditions through simulation. Instead of physically testing each condition, the patent uses computational models to generate and compare simulated resist images, allowing rapid evaluation of multiple process conditions without the time cost of actual experimental measurements or iterations.
4Manufacturing precision
If conventional OPC is used without process window consideration, then the mask error enhancement factor is minimized at nominal condition, but edge placement errors increase under process variation
Solution Approach 1:
The patent segments the edge placement error analysis into multiple process condition categories (different focus offsets and exposure doses). By evaluating and correcting edge placement separately for each segment, the system addresses the harmful effects of process variation systematically, ensuring that edge placement accuracy is maintained across all conditions rather than degrading under variation.
Solution Approach 2:
The patent applies preliminary anti-action by pre-calculating OPC corrections that specifically counteract expected edge placement errors under various process conditions. The system identifies potential harmful effects (focus and exposure variations) in advance and applies corrective measures to compensate for them, preventing edge placement errors before they occur during actual lithography processing.
Data Source
AI summary
One embodiment of a method for process window optimized optical proximity correction includes applying optical proximity corrections to a design layout, simulating a lithography process using the post-OPC layout and models of the lithography process at a plurality of process conditions to produce a plurality of simulated resist images. A weighted average error in the critical dimension or other contour metric for each edge segment of each feature in the design layout is determined, wherein the weighted average error is an offset between the contour metric at each process condition and the contour metric at nominal condition averaged over the plurality of process conditions. A retarget value for the contour metric for each edge segment is determined using the weighted average error and applied to the design layout prior to applying further optical proximity corrections.