Semiconductor Defect Coordinate Mapping Using Golden Image Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing semiconductor fault detection methods, such as optical fault isolation (OFI), suffer from low accuracy and long detection times due to errors in coordinate alignment and worker-dependent analysis, which can exceed 50 micrometers and take up to ten minutes per device.
Innovation Solution
An apparatus and method that automatically extracts fault coordinates by generating a golden image from pattern images, creating a design layout image, and using relationship information to accurately align defect positions, reducing human intervention and improving detection speed and precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If optical fault isolation (OFI) equipment is used to detect fault coordinates, then fault detection can be performed nondestructively, but the accuracy is low with errors up to 50 micrometers
Solution Approach 1:
The patent creates a golden image by copying and averaging patterns from multiple normal semiconductor devices. This golden image serves as a reference template that captures the ideal pattern structure, enabling precise comparison and fault detection without the coordinate alignment errors of direct OFI measurement
Solution Approach 2:
The system performs preliminary actions by pre-acquiring pattern images from multiple normal devices, generating the golden image and establishing the coordinate transformation relationship with the design layout before actual fault detection. This preparation eliminates the need for real-time coordinate alignment during fault detection, improving both accuracy and speed
2Reliability
If a worker manually compares pattern images with design layout to extract fault coordinates, then fault detection can be performed, but the detection time is long averaging ten minutes per device
Solution Approach 1:
The patent replaces the manual mechanical comparison process with an automated image processing system. The processor automatically compares the defect reaction image against the golden image and design layout, extracting fault coordinates through algorithmic coordinate transformation rather than human visual inspection, thereby dramatically increasing detection speed
Solution Approach 2:
The golden image acts as an intermediary between the raw pattern images and the design layout. It serves as a processed reference that bridges the optical measurement domain and the design coordinate system, enabling automated and rapid fault coordinate extraction without direct manual intervention
3Reliability
If a worker manually extracts fault coordinates by comparing images, then fault detection can be performed, but the accuracy differs for each worker
Solution Approach 1:
The system performs self-service by automatically completing the entire fault detection process including image comparison, coordinate extraction, and fault identification. The automated processor eliminates human variability entirely, ensuring consistent and reproducible results across all detections without relying on individual worker skill levels
Data Source
AI summary
Apparatus for analyzing images includes a memory and a processor, and the processor is configured to, by executing the at least one instruction stored in the memory, detect an image coordinate of a defect position in a defect reaction image acquired by imaging a semiconductor device, generate a golden image of the semiconductor device based on a plurality of pattern images acquired by imaging a plurality of semiconductor devices, generate a design layout image corresponding to the pattern image based on a design layout of the plurality of semiconductor devices, generate relationship information between a pattern image and the design layout image based on the pattern image, the golden image, and the design layout image, and extract a design layout image coordinate corresponding to the defect position from the design layout image using the image coordinate of the defect position in the defect reaction image and the generated relationship information.