Adaptive Sampling for Semiconductor Inspection Without Design Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Semiconductor manufacturing faces challenges in detecting defects due to tool noise and image misalignment, especially in tight process windows, where design layout information is not readily available, making it difficult to identify critical regions for inspection.
Innovation Solution
A computer-implemented method for adaptive sampling using a high-speed, low-resolution tool to generate images, segment them, determine feature vectors, cluster, and select samples for higher-resolution tool imaging, allowing for efficient identification of critical regions and defect classification without relying on design data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution tool is used to inspect all locations on wafer, then measurement precision is improved, but productivity deteriorates due to time-consuming inspection process
Solution Approach 1:
The wafer surface is segmented into multiple locations based on clustering analysis of image features. The inspection process is divided into two stages: first, a high-speed tool captures images at all locations and segments them into groups; second, a high-resolution tool only inspects selected representative locations from each cluster. This segmentation allows the system to maintain high precision for critical areas while improving overall productivity by avoiding redundant high-resolution inspections at all locations.
Solution Approach 2:
Different inspection resolutions are applied to different locations on the wafer based on their criticality. The system identifies critical regions through feature vector clustering and applies high-resolution inspection only to locations that cluster together as critical, while using high-speed inspection for non-critical areas. This local quality approach ensures measurement precision is maintained where needed without sacrificing productivity across the entire wafer.
2Ease of operation
If design layout information is used to identify critical regions, then ease of operation is improved, but adaptability deteriorates when design data is unavailable
Solution Approach 1:
The system performs self-service by automatically extracting critical region information directly from the captured images through feature vector calculation and clustering algorithms. Instead of relying on external design layout data, the system analyzes image features such as geometric patterns, material properties, and structural characteristics to automatically identify critical regions. This self-service capability maintains ease of operation while significantly improving adaptability to situations where design data is unavailable or inaccurate.
Solution Approach 2:
The patent replaces the mechanical/reliant approach of using pre-existing design layout information with an automated computational approach using image processing and machine learning algorithms. The system substitutes the need for design data access with direct analysis of actual wafer images, using clustering algorithms to identify critical regions based on visual features rather than relying on stored design information. This substitution enhances adaptability while maintaining operational ease.
Data Source
AI summary
Methods and systems for adaptive sampling for semiconductor inspection recipe creation, defect review, and metrology are provided. The embodiments provide image processing and pattern recognition algorithms and an adaptive sampling method for extracting critical areas from SEM image patches for use in a wafer inspection system when design data for a semiconductor chip is not available. The embodiments also provide image processing and pattern recognition algorithms for efficiently discovering critical defects and significant deviations in the normal manufacturing process, using the output from a wafer inspection system and an adaptive sampling method to select wafer locations to be examined on a high resolution review or metrology tool.