Wafer Alignment Recipe Using Multi-Site Characteristic Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor manufacturing, conventional wafer alignment systems face challenges due to significant changes in characteristic patterns during processes like chemical mechanical polishing, leading to accuracy issues in wafer placement and inspection.
Innovation Solution
A method and system for generating and using multiple alignment data sets, including image information and coordinates of characteristic points at different sites on a wafer, to create a recipe for alignment, which allows for accurate identification and updating of alignment data sets to account for pattern changes, thereby improving precision in wafer alignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional single-site alignment is used, then the alignment process is simple, but alignment accuracy deteriorates when characteristic patterns change during manufacturing
Solution Approach 1:
The patent divides the alignment task into multiple independent alignment data sets, each corresponding to a different site on the wafer. Each data set contains image information and characteristic point coordinates for a specific site. By segmenting the alignment process into multiple sites rather than relying on a single site, the system maintains alignment accuracy even when characteristic patterns change at any individual site during manufacturing processes like chemical mechanical polishing.
Solution Approach 2:
The patent performs preliminary alignment at multiple sites before the actual wafer processing and inspection. Multiple alignment data sets are captured and stored in advance, creating a reference framework that accounts for potential pattern changes. This preliminary multi-site alignment establishes a robust coordinate system that can accommodate subsequent pattern variations without compromising alignment accuracy.
2Measurement precision
If multiple alignment data sets are used, then alignment accuracy improves, but data processing complexity increases
Solution Approach 1:
The patent implements a feedback mechanism where the system automatically identifies which alignment data set corresponds to the current wafer site by matching image information. The system compares captured wafer images with stored alignment data set images, identifies the matching data set, and retrieves the corresponding characteristic point coordinates. This automated feedback loop eliminates manual data management complexity while maintaining high alignment accuracy through multiple data sets.
Solution Approach 2:
The alignment system performs self-service by automatically selecting and applying the appropriate alignment data set based on real-time image recognition. The system autonomously matches the current wafer site with the corresponding pre-captured alignment data, retrieves coordinates, and performs alignment without requiring manual intervention. This self-service capability manages the complexity of multiple data sets while maintaining ease of operation.
3Ease of operation
If characteristic patterns are used for alignment, then alignment is straightforward, but patterns may not be identifiable after processing steps like chemical mechanical polishing
Solution Approach 1:
The patent captures and stores alignment data including image information and characteristic point coordinates at multiple sites on the wafer before any processing steps are applied. This preliminary action creates a reference database of pre-processing wafer states. When the wafer undergoes processing like chemical mechanical polishing that may alter characteristic patterns, the system can still perform accurate alignment by comparing post-processing images with the pre-captured alignment data sets, ensuring both ease of operation and reliability.
Solution Approach 2:
The patent creates optical copies (images) of the wafer surface at multiple sites before processing, storing these as alignment data sets. These image copies serve as reference templates that can be compared against the actual wafer surface even after processing-induced pattern changes. By working with these copied reference images rather than relying solely on the physical characteristic patterns that may change, the system maintains both operational simplicity and pattern identifiability reliability.
Data Source
AI summary
A method for a system to generate a recipe for performing wafer alignment, includes: generating first and second alignment data sets, the first alignment data set including image information regarding a first site on a wafer and coordinates of characteristic points at the first site, and the second alignment data set including image information regarding a second site different than the first site on the wafer and coordinates of characteristic points at the second site; and saving the generated first and second alignment data sets as a recipe for wafer alignment; wherein the generating of the first alignment data set includes: selecting, as the first site, a site including a characteristic pattern on the wafer; determining first and second characteristic points at the selected site; and recording coordinates of the determined first and second characteristic points.