Semiconductor Cell Counting via Stage Motion Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Semiconductor miniaturization has made it difficult to accurately locate and count failed cells with existing methods, as the accuracy of stopping stages is compromised, leading to erroneous recognition and impractical counting of cells due to unstable stage movement and pattern recognition limitations.
Innovation Solution
A semiconductor testing method using a charged particle beam to display cells on a monitor, with a movable stage controlled to correct errors and count cells quickly and accurately, employing rectangular frames and numeric values for precise identification and counting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the stage is moved for each screen to enable pattern recognition, then image recognition can be performed, but the stopping accuracy becomes compromised and cells may be erroneously detected
Solution Approach 1:
The patent replaces the mechanical stage stopping method with a computational approach. Instead of relying on the mechanical stage to stop at precise positions, the system captures images at multiple overlapping positions and uses image recognition algorithms to identify cells. This substitution of mechanical precision with computational processing resolves the contradiction between stage stopping accuracy and cell detection accuracy.
Solution Approach 2:
The patent performs preliminary actions by capturing images at multiple positions before final cell identification. The system takes overlapping images at different stage positions, processes them through pattern recognition, and then identifies cells based on the combined information. This preliminary image capture and processing enables accurate cell detection without requiring precise mechanical stopping.
2Measurement precision
If the stage is stopped for each screen to correct position, then pattern matching can be performed, but the counting speed becomes slow
Solution Approach 1:
The patent implements continuous useful action by moving the stage continuously through the sample while capturing images at multiple overlapping positions, rather than stopping at each position. The image recognition process runs continuously on the captured images, enabling both accurate pattern recognition and high-speed cell counting simultaneously.
Solution Approach 2:
The patent adds the time dimension to the imaging process by capturing multiple images at different time points during continuous stage movement. This temporal dimension allows the system to process overlapping images computationally, achieving both accuracy and speed without mechanical stopping.
3Extent of automation
If pattern recognition is used to identify cells, then cell identification can be automated, but the initial and stopping speeds become unstable
Solution Approach 1:
The patent replaces mechanical speed control with computational image processing. Instead of attempting to maintain stable stage speeds mechanically, the system captures images during stage movement and uses pattern recognition algorithms to identify cells. The automation of cell identification through image processing compensates for the instability in mechanical speed control.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables rapid and accurate counting of semiconductor cells, improving usability by stabilizing stage movement and reducing errors in cell recognition, allowing for precise location and measurement of failures.
Implementation Method 1
a plurality of cells formed in a semiconductor are displayed on display means based on a sample signal obtained by irradiating the semiconductor on a movable stage with a charged particle beam
Data Source
AI summary
A semiconductor testing method capable of quickly counting semiconductor cells in which a seemingly horizontal or vertical line is drawn with a mouse, and raster rotation is performed in alignment with the closer axis. After that, the stage is horizontally moved, pattern matching is performed on an image on a position where the image should be disposed, and an angle is adjusted. The stage is moved evenly along the X-axis and the Y-axis, achieving a movement to a destination like a straight line. In synchronization with the smooth movement of the stage, a cell is surrounded in a rectangular frame by a ruler, and the number of cells is displayed with a numeric value.