Semiconductor Mask Defect Inspection via Segmented PSF
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional defect inspection methods for semiconductor masks, such as die-to-database inspection, face challenges in accurately reproducing optical blur, leading to increased probabilities of erroneous defect detection due to difficulties in reproducing optical blur using convolution operations that involve multiple edge portions of circuit patterns.
Innovation Solution
A defect inspection device and method that acquires and compares images using a processor to generate a reference image based on correct data from a database, estimating defective pixels by comparing the reference image with a captured image, thereby improving accuracy in defect detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If convolution operation is used to reproduce optical blur, then the reference image can be generated, but the accuracy of reproducing optical blur decreases due to involvement of multiple edge portions
Solution Approach 1:
The patent divides the circuit pattern into multiple regions based on edge orientation (e.g., horizontal edges, vertical edges, diagonal edges). For each region, a separate PSF is extracted that corresponds to that specific edge orientation. This segmentation allows the system to avoid mixing different edge portions in a single PSF, thereby accurately reproducing optical blur for each specific region without contamination from other edge orientations.
Solution Approach 2:
The patent applies different PSFs to different regions of the circuit pattern based on their local edge characteristics. Each region receives a PSF that is specifically tailored to its edge orientation and local optical characteristics. This local quality approach ensures that the optical blur reproduction is optimized for each specific region rather than using a single universal PSF, thereby improving both measurement precision and reliability.
2Ease of manufacture
If a single PSF is used for the entire circuit pattern, then the processing is simplified, but the optical blur reproduction accuracy decreases
Solution Approach 1:
The patent segments the circuit pattern into multiple regions with distinct edge orientations and extracts a separate PSF for each region. This segmentation maintains processing efficiency by using automated region classification algorithms while significantly improving optical blur reproduction accuracy by matching PSFs to specific edge characteristics rather than using a single generic PSF for the entire pattern.
Solution Approach 2:
The patent changes the parameters of the PSF based on the local characteristics of different circuit pattern regions. Specifically, it adjusts the PSF parameters (such as blur amount and orientation) to match the edge characteristics of each region. This parameter adaptation allows the system to maintain processing efficiency while achieving high-accuracy optical blur reproduction for diverse pattern features.
Data Source
AI summary
A defect inspection device includes at least one memory storing instructions and at least one processor. The at least one processor is configured to execute the instructions to acquire a first image of an inspection target created in a first creation method, acquire a second image obtained by photographing the inspection target, extract index data similar to the acquired first image with reference to a database, the index data being a third image created in the first creation method or a feature quantity obtained by the third image, the database including the index data associated with correct data that is used as a comparison target of the index data and is an image determined not to be defective in previous inspection, acquire the correct data associated with the extracted index data in the database, generate a reference image on the basis of the acquired correct data, and estimate a pixel associated with a defective position on the inspection target photographed in the second image by comparing the reference image with the second image.