Lithographic Stitching Error Analysis via Distortion Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Lithographic patterning processes face challenges in accurately positioning and aligning features across different regions on a substrate, leading to stitching errors that affect the quality and yield of the patterning process, especially at the boundaries between non-overlapping regions.
Innovation Solution
An apparatus and method that utilize one or more processors to analyze images of substrates with features from multiple lithographic exposures, determining performance metrics such as distance, magnification, and deformation errors by comparing feature characteristics at the boundaries between regions, using machine learning models and vision technology to identify stitching errors and optimize the patterning process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If multiple masks are moved across regions of a substrate to pattern different portions of the pattern sequentially, then the pattern can be formed on the substrate, but stitching errors occur at the boundaries between regions
Solution Approach 1:
Alignment marks are incorporated into the pattern design before lithographic exposure. These marks are formed simultaneously with the pattern features in the same lithographic step, establishing reference points that will be used for subsequent alignment operations. This preliminary action ensures that alignment references are available and precisely positioned before the multi-step patterning process begins.
Solution Approach 2:
Alignment marks serve as intermediary elements between different lithographic exposures. These marks are not part of the final product pattern but function as mediators that enable precise registration between multiple patterning steps. The alignment marks are detected and used to calculate transformation parameters that guide the positioning of subsequent masks, thereby mediating the alignment process between different exposure regions.
2Productivity
If the mask is moved across the substrate to expose the same pattern multiple times, then the pattern can be formed on the substrate, but positioning errors occur between exposed portions
Solution Approach 1:
The system detects the actual positions of alignment marks after each lithographic exposure and uses this information to calculate correction parameters for subsequent exposures. This feedback loop allows the system to compensate for positioning errors that occur during mask movement, continuously refining the alignment accuracy across multiple patterning steps based on actual measured positions rather than relying solely on theoretical positioning.
Solution Approach 2:
The patent replaces reliance on purely mechanical positioning systems with an optical/detector-based alignment system. Instead of depending solely on the precision of mechanical mask movement and positioning mechanisms, the system uses optical detection of alignment marks to determine actual positions and calculate corrections, substituting mechanical precision requirements with optical measurement and computational compensation.
3Manufacturing precision
If sophisticated fine-tuning steps are applied to optimize NA and use phase shifting patterning devices, then pattern reproduction quality improves, but process complexity increases
Solution Approach 1:
The alignment and measurement functions are extracted from the main lithographic exposure process. Dedicated alignment marks and detection mechanisms are separated from the product pattern formation, allowing the exposure system to focus on high-quality pattern reproduction while the alignment system independently handles positioning measurements and corrections without interfering with the primary patterning function.
Data Source
AI summary
A method for characterizing a patterning process, the method including obtaining a plurality of values of stitching errors made along one or more boundaries between at least two patterned adjacent fields or sub-fields on a substrate; and fitting, using a hardware computer system, a distortion model to the plurality of values to obtain a fingerprint representing deformation of a field or sub-field out of the at least two patterned adjacent fields or sub-fields.