Wafer Alignment Mark Recognition Using Binary Reference Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing semiconductor manufacturing processes face challenges in accurately aligning to patterns on wafers for efficient inline measurement, leading to reduced production efficiency and yield.
Innovation Solution
A method involving the use of a high contrast pixelated black-and-white image as a reference to recognize alignment mark patterns on wafers, enabling precise alignment and measurement by generating a synthetic image in black and white pixels for improved recognition rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional image recognition methods are used to identify alignment mark patterns on wafers, then the process can be performed with standard equipment, but the recognition rate is low leading to reduced measurement efficiency
Solution Approach 1:
The patent converts the alignment mark pattern image into a binary image with only two grayscale levels (black and white), creating high contrast between the alignment mark and background. This binary conversion enhances the visual distinction and enables more reliable automated recognition by the metrology system, directly improving both recognition rate and measurement efficiency.
2Productivity
If inline measurement steps are inserted to monitor patterns on wafers, then real-time measurement data can be obtained, but the alignment accuracy to the pattern is insufficient reducing production efficiency
Solution Approach 1:
The patent performs preliminary image processing to generate a binary reference image of the alignment mark pattern before conducting the actual measurement. This pre-processing step creates a standardized, high-contrast reference that facilitates accurate alignment during the measurement process, ensuring both high alignment accuracy and maintained production efficiency through automated processing.
3Ease of operation
If standard image processing is used for alignment mark recognition, then the system complexity remains low, but the recognition rate is insufficient increasing the burden of inline process control
Solution Approach 1:
The patent changes the grayscale parameter of the image by converting it to a binary image with only two grayscale levels. This parameter transformation simplifies the image data while dramatically improving the recognition rate of the alignment mark pattern, reducing the burden of inline process control through more reliable automated recognition without significantly increasing system complexity.
Data Source
AI summary
A method for aligning to a pattern on a wafer is disclosed. The method includes the steps of obtaining a first inline image from a first sample wafer, obtaining a first contour pattern of an alignment mark pattern from the first inline image, using the first contour pattern to generate a first synthetic image in black and white pixels of only two grayscale levels, using the first synthetic image as a reference to recognize the alignment mark pattern on a tested wafer, and aligning to a tested pattern on the tested wafer according to a position of the alignment mark pattern on the tested wafer and a coordinate information.