Semiconductor Wafer Defect Diagnosis Using Multi-Image Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor manufacturing, existing inspection methods fail to accurately detect defects in wafers as dimensions decrease, leading to quality and reliability issues due to smaller defects causing aberrations and yield losses.
Innovation Solution
A method involving an inspection apparatus that captures multiple raw images from semiconductor wafers using GDS files, performs first and second image-based comparisons to detect and classify defects, utilizing a reference image from a golden wafer or die to verify and diagnose defects, thereby improving defect detection accuracy and reducing tooling time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing inspection methods are used, then manufacturing process is simple, but defect detection accuracy deteriorates as dimensions decrease
Solution Approach 1:
The inspection method is divided into multiple sequential stages: obtaining reference images from golden wafers, capturing raw images of production wafers, performing first image-based comparison to detect differences, and conducting second image-based comparison to classify defect types. This segmentation allows each stage to focus on specific detection tasks, improving overall accuracy without overwhelming complexity
Solution Approach 2:
Reference images are obtained from golden wafers before inspection begins. These reference images serve as baseline standards for comparison, allowing the system to pre-establish what constitutes normal versus defective patterns. This preliminary action enables more accurate defect detection during subsequent inspections
2Measurement precision
If multiple image comparisons are performed, then defect detection accuracy improves, but inspection time increases
Solution Approach 1:
The comparison process is segmented into two distinct phases: first image-based comparison for detecting image differences, and second image-based comparison for classifying defect types. This segmentation allows the system to perform comprehensive analysis while organizing operations efficiently, reducing unnecessary processing time
Solution Approach 2:
The system performs comparisons on specific regions of interest identified through the first comparison, rather than analyzing entire wafer surfaces uniformly. By focusing computational resources on areas showing differences, the system achieves high accuracy while minimizing total inspection time
3Measurement precision
If reference images from golden wafers are used, then defect classification accuracy improves, but system complexity increases
Solution Approach 1:
Reference images are obtained by capturing images from golden wafers that represent defect-free or known-good standards. These copied reference images are stored and used for comparison against production wafers, providing an accurate baseline without requiring complex real-time reference generation systems
Solution Approach 2:
Golden wafers are prepared and reference images are captured in advance, before the actual inspection process begins. This preliminary preparation creates a library of reference standards that simplifies the main inspection system, as it only needs to compare against pre-established references rather than generate them dynamically
Data Source
AI summary
Methods and systems for diagnosing a semiconductor wafer are provided. A plurality of raw images are captured with a tilt angle from the semiconductor wafer according to graphic data system (GDS) information regarding a layout of a target die, by an inspection apparatus. A first image-based comparison is performed on the plurality of raw images, by a determining circuitry, to obtain a defect image in the plurality of raw images. A second image-based comparison is performed on a reference image and the defect image, so as to classify a defect type of an image difference in the defect image, by the determining circuitry. The number of the plurality of raw images is greater than 2.