Systematic Defect Detection in Semiconductor Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional defect review methods in semiconductor manufacturing are insufficient for efficiently detecting systematic defects, especially as the number of hotspots to monitor increases, leading to potential overlooks in process fluctuations and reduced production yield.
Innovation Solution
A defect review method and apparatus that utilize scanning electron microscope (SEM) images to classify candidate defects, computing feature values, and comparing positional information from design data to identify systematic defects caused by pattern layout or equipment properties, enabling efficient detection of systematic defects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional defect review methods are used to monitor multiple hotspots, then the coverage of defect detection is improved, but the complexity of the review process increases and systematic defects may be overlooked
Solution Approach 1:
The patent segments the defect review process into distinct functional modules: image acquisition unit, image processing unit, defect detection unit, and systematic defect determination unit. Each module handles a specific aspect of the review process, allowing complex multi-hotspot monitoring to be broken down into manageable steps that can be executed systematically without overwhelming the review system
Solution Approach 2:
The patent performs preliminary actions by acquiring design data and hot spot coordinate data before the actual defect review process. The systematic defect determination unit compares detected defects against pre-stored hot spot coordinates to identify systematic defects. This preliminary preparation enables efficient processing during mass production without increasing real-time review complexity
2Reliability
If the number of hotspots to be monitored increases, then the detection coverage is improved, but the probability of overlooking shape defects increases
Solution Approach 1:
The patent implements feedback mechanisms where the systematic defect determination unit continuously compares detected defect positions against stored hot spot coordinate data. This feedback loop ensures that even as the number of monitored hotspots increases, each defect is systematically evaluated against known critical locations, maintaining high detection accuracy without missing shape defects amidst the increased coverage
Solution Approach 2:
The system performs preliminary identification of hot spots (critical areas prone to systematic defects) before mass production based on design data and exposure simulation results. By pre-mapping these critical zones and storing their coordinates, the system can quickly compare detected defects against these known high-risk locations, ensuring precise shape defect detection even when monitoring a large number of hotspots simultaneously
3Measurement precision
If manual review of each defect is performed, then the accuracy of defect classification is improved, but the productivity of the review process decreases
Solution Approach 1:
The patent implements self-service automation where the defect review system automatically performs image processing, defect detection, classification, and systematic defect determination without requiring manual intervention for each defect. The system uses automated image processing techniques to extract defect features, compares them against stored hot spot data, and independently determines systematic defects, thereby maintaining high classification accuracy while dramatically improving review throughput and productivity
Solution Approach 2:
The patent replaces manual mechanical review processes with automated electronic systems. Image processing units automatically analyze defect images, detection algorithms automatically identify and classify defects, and systematic defect determination is automatically performed by comparing coordinates. This substitution of manual operations with automated computational processes maintains precision while enabling high-speed processing of multiple hotspots
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The method allows for accurate and efficient detection of systematic defects, improving the accuracy of defect classification and reducing the likelihood of overlooking shape defects, thereby enhancing production yield and process monitoring.
Implementation Method 1
an electron beam 115 is emitted from an electron gun 115 toward a sample 114
Implementation Method 2
electrons 1152 (secondary or reflective electrons) emitted therefrom are acquired or detected by a detector 116
Data Source
AI summary
A candidate-defect classification method includes the steps of acquiring a scanning electron microscope (SEM) image of a candidate defect detected in an inspection from a sample including a pattern formed thereon, the inspection being preliminarily performed by an other inspection device; computing a feature value of the candidate defect by processing the SEM image; executing defect classification of the candidate defect as any one of a pattern shape defect and an other defect by using the computed feature value; acquiring positional information contained in design data of the pattern with respect to a candidate defect classified as the pattern shape defect; and extracting a systematic defect from among candidate defects classified as the pattern shape defects by performing a comparison of the positional information contained in the design data of the acquired candidate defect to positional information of a portion that has a high probability of causing a pattern formation failure and that has been obtained from the design data of the pattern, the systematic defect being caused due to a layout shape of the pattern, properties of a processor for forming the pattern, or the like.