Rotational Center Determination for Alignment Marks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the center of rotationally symmetrical alignment marks in semiconductor manufacturing are prone to errors due to manual definition of reference points, leading to misalignment issues between substrates like masks and wafers, especially when they have different dimensions caused by thermal expansion.
Innovation Solution
An automated method using image recognition software to determine the rotational center of rotationally symmetrical alignment marks by training the software to recognize alignment marks in multiple orientations, allowing for sub-pixel resolution alignment, and using trigonometric analysis to pinpoint the central point of the alignment mark.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual definition of reference points is used to determine alignment mark centers, then the process is simple to implement, but alignment accuracy deteriorates due to manual errors
Solution Approach 1:
The patent replaces manual mechanical reference point definition with an automated image recognition system. The software automatically detects alignment marks in multiple orientations, determines their centers through trigonometric analysis, and calculates rotational centers without manual intervention. This substitution of manual mechanical operations with automated optical and computational methods resolves the contradiction by improving measurement precision while accepting increased process complexity.
Solution Approach 2:
The image recognition software performs self-training to automatically recognize alignment marks and their orientations. The system independently identifies marks in multiple orientations, calculates rotational centers through trigonometric analysis, and determines reference points without requiring manual definition. This self-service capability improves alignment accuracy by eliminating manual errors while automating the previously simple process.
2Adaptability or versatility
If independent training sequences are used for mask and wafer alignment marks, then each mark can be recognized independently, but the number of training steps increases
Solution Approach 1:
The patent performs preliminary action by training the image recognition software to recognize alignment marks in multiple orientations (0°, 90°, 180°, 270°) before actual alignment operations. This preliminary multi-orientation training enables the system to quickly and accurately recognize marks regardless of their rotational position during alignment, reducing the time needed for mark identification while maintaining flexibility in handling different mark geometries and orientations.
3Ease of operation
If reference points are defined away from the center of alignment marks, then manual definition is easier, but alignment precision deteriorates
Solution Approach 1:
The patent replaces manual reference point definition with automated image recognition and trigonometric calculation. The system automatically identifies the center of alignment marks by analyzing their geometry and calculating rotational centers through trigonometric relationships from multiple orientations. This eliminates the need for manual reference point definition while achieving higher precision by accurately determining centers and rotational positions through computational methods.
Data Source
AI summary
The invention provides a method for determining the center of a rotationally symmetrical alignment mark thereby using an image recognition software. The alignment mark is recognized by the image recognition software in different orientations by rotating the alignment mark around a symmetry angle with respect to which the alignment mark is rotationally symmetrical, and respective reference points are determined. The center of rotation of the determined reference points corresponds to the center of the alignment mark. Moreover, the invention provides a method for aligning two flat substrates each having a rotationally symmetrical alignment mark and being arranged substantially parallel with respect to each other. To this end, the centers of the alignment marks of the two substrates are determined by means of the method for determining the centers according to the present invention, and the two substrates are aligned by moving at least one of the two substrates in parallel so that the positions of the centers of the alignment marks coincide.