Super-Resolution Image Reconstruction for Semiconductor Wafer Defect Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Semiconductor wafer inspection systems face challenges in detecting defects due to the limitations of resolution, leading to blurry images and reduced signal-to-noise ratios, which hinder the detection of features smaller than the inspection system's capabilities.
Innovation Solution
The implementation of a multi-mode inspection system that includes an illumination sub-system and a collection sub-system, utilizing super-resolution image reconstruction techniques such as deconvolution and regularization to enhance image resolution and detect defects, by estimating the point spread function and applying iterative deconvolution methods like Richardson-Lucy deconvolution with regularization terms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional optical inspection systems are used to inspect semiconductor wafers, then inspection speed and coverage can be maintained, but resolution is insufficient leading to blurry images and reduced signal-to-noise ratio for defect detection
Solution Approach 1:
The patent applies preliminary action by performing image reconstruction and super-resolution processing on acquired wafer images before defect detection. The system reconstructs multiple images taken at different focal depths or angles, enhances their resolution using algorithms like deconvolution or iterative reconstruction, and then performs defect detection on the enhanced images. This preliminary enhancement of image quality before analysis improves both resolution and signal-to-noise ratio for defect detection.
2Measurement precision
If image resolution is increased to detect smaller features, then defect detection capability improves, but inspection complexity and processing requirements increase
Solution Approach 1:
The patent applies dimensionality change by acquiring images at multiple focal depths (z-stacking) or multiple angles and then reconstructing a high-resolution 3D or super-resolution 2D image. Instead of attempting to achieve high resolution in a single 2D image, the system utilizes the additional depth or angular dimension to gather more information, which is then synthesized into an enhanced image with superior resolution and signal-to-noise ratio through computational reconstruction.
Data Source
AI summary
An inspection system includes an illumination sub-system, a collection sub-system, and a controller. The illumination sub-system includes an illumination source configured to generate a beam of illumination and a set of illumination optics to direct the beam of illumination to a sample. The collection sub-system includes a set of collection optics to collect illumination emanating from the sample and a detector configured to receive the collected illumination from the sample. The controller is configured to acquire a test image of the sample, reconstruct the test image to enhance the resolution of the test image, and detect one or more defects on the sample based on the reconstructed test image.