QLIVBM Image Alignment Using Independent Vector Targets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image alignment methods for semiconductor manufacturing face challenges such as false alignment and alignment failure due to sensitivity issues, leading to interruptions and reduced manufacturing yield, especially when alignment sites are non-optimal, damaged, or missing.
Innovation Solution
A quantitative linear independent vector based method (QLIVBM) that aligns candidate alignment target images to setup images using mutually linearly independent vectors, reducing false alignment and ensuring accurate alignment even with non-optimal sites.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If alignment algorithms are made less sensitive to alignment site image variations, then the chance of achieving alignment increases, but false alignment occurs and premap and PDA interruption is caused
Solution Approach 1:
The patent divides the alignment process into multiple independent alignment sites (first alignment site and second alignment site) rather than relying on a single alignment site. Each alignment site is processed separately through its own alignment algorithm, and the results are combined to determine the final alignment. This segmentation reduces false alignment by ensuring that if one site fails or produces incorrect results, the other site can provide corrective information.
Solution Approach 2:
The patent introduces a reference alignment site that serves as an intermediary element. The reference alignment site is used to establish a known correct alignment position, and alignment algorithms are developed that specifically target this reference site to verify and correct alignment accuracy. This intermediary reference point allows the system to detect and correct false alignment without requiring multiple complex algorithms to compete.
2Manufacturing precision
If alignment algorithms are made more sensitive to alignment site image variations, then false alignment is reduced, but alignment fails frequently due to process change causing instant alignment interruption
Solution Approach 1:
By segmenting the alignment process into multiple independent alignment sites, the patent ensures that sensitivity at one site does not cause overall alignment failure. If one alignment site is affected by process changes or variations, the other alignment sites can compensate, maintaining continuous alignment operation without interruption.
Solution Approach 2:
The patent prepares for alignment failures by establishing multiple alignment sites beforehand. This cushioning approach ensures that if one alignment site fails due to sensitivity issues or process changes, the system has pre-prepared alternative alignment sites that can immediately take over, preventing alignment interruption.
3Device complexity
If a single alignment site is used for wafer alignment, then the setup phase is simplified, but false alignment and low die corner accuracy determination are possible
Solution Approach 1:
The patent segments the alignment site into multiple locations (first alignment site and second alignment site) with distinct patterns. This segmentation maintains relative simplicity in the setup phase while significantly improving die corner accuracy by providing multiple reference points for alignment calculation, eliminating the false alignment possibilities inherent in single-site alignment.
Data Source
AI summary
Methods and systems for image alignment are provided. One method includes separately aligning candidate alignment target images in images generated for a specimen to corresponding setup images for setup alignment targets selected to have mutually linearly independent vectors between locations of the setup alignment targets and a reference location. The method also includes, for any of the candidate images successfully aligned to its corresponding setup image, separately determining coordinates of the reference location from coordinates of the aligned candidate images and their corresponding mutually linearly independent vectors. In addition, the method includes determining final coordinates of the reference location in the images generated for the specimen from the separately determined coordinates of the reference location.