Lithographic Edge Placement Predictor Using Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current photolithography technologies face challenges in accurately predicting and monitoring the edge placement of image patterns on semiconductor wafers, leading to potential malfunctions and reduced yield due to spatial misplacement of image patterns across different layers, and lack methods to optimize the projection optics system for minimizing pattern overlay errors.
Innovation Solution
A method and system that involve performing wafer-exposure runs to determine operational factors, adjusting optical components of the projection system, and using a computer device to calculate and correct the edge positions based on optical imaging parameters and shot-overlay characteristics, allowing for precise alignment and optimization of image placement without modifying scanning synchronization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If photolithography is used to define features during semiconductor wafer processing, then patterns can be formed on the wafer, but spatial misplacement of image patterns across different layers occurs leading to degradation of integrated device operation
Solution Approach 1:
The patent applies preliminary action by predicting pattern edge positions before actual lithographic exposure using a trained neural network model. The system calculates expected edge placements based on process parameters and reticle design data in advance, allowing for pre-correction of potential misplacement issues before they affect the final device operation
Solution Approach 2:
The patent implements feedback by monitoring actual pattern edge positions after lithographic exposure and comparing them with predicted positions. The system uses measured deviations to update and refine the neural network model, creating a closed-loop control system that continuously improves placement accuracy across production batches
2Measurement precision
If traditional lithographic monitoring methods are used, then pattern formation can be tracked, but edge placement positions cannot be predicted with sufficient precision
Solution Approach 1:
The patent introduces an intermediary neural network model that acts as a mediator between process parameters and actual edge placements. This neural network translates input parameters (exposure conditions, focus settings, reticle data) into predicted edge positions, bridging the gap between controllable process variables and difficult-to-measure placement outcomes
Solution Approach 2:
The patent replaces traditional mechanical measurement systems with a computational approach using neural networks. Instead of relying solely on physical measurement tools to detect edge positions, the system uses machine learning algorithms to predict placements based on process data, enabling higher precision without additional measurement hardware
3Manufacturing precision
If the projection optical system is adjusted to correct edge placement, then pattern alignment improves, but system complexity increases
Solution Approach 1:
The patent applies preliminary action by calculating optimal optical system adjustments in advance based on predicted edge placement deviations. The system determines necessary corrections to projection optics parameters before actual exposure, allowing for proactive alignment correction rather than reactive tuning during production
Solution Approach 2:
The patent implements feedback by using measured edge placement deviations to automatically adjust projection optical system parameters. The system creates a closed-loop control mechanism where placement errors directly trigger optical corrections, reducing the need for complex manual intervention and operator expertise
Data Source
AI summary
Method and system configured to reduce or even nullify the degradation of images created by the projector tool turns on the optimization of the pattern-imaging by adjusting parameters and hardware of the projector to judiciously impact the placement of various image edges at different locations in the image field. Adjustments to the projector (exposure tool) include a change of a setup parameter of the exposure tool and/or scanning synchronization and/or a change of a signature of the optical system of the exposure tool determined as a result of minimizing the pre-determined cost function(s) that are parts of a comprehensive edge-placement error model.