Hierarchical Clustering of Fourier Transform Layout Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In the manufacturing of integrated circuits, existing inspection systems face challenges in efficiently and accurately analyzing the massive data from SEM images due to the complexity and large number of features, leading to time-consuming pattern grouping and clustering processes with predefined, fixed parameters that lack user control.
Innovation Solution
The proposed system and method involve receiving image data of patterns on a wafer, performing a Fourier Transform to separate patterns into sets, and applying hierarchical clustering to group patterns based on similarity, allowing for recursive partitioning and user-adjustable parameters to customize the clustering process, thereby improving efficiency and accuracy in hotspot analysis and wafer inspection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If traditional pattern grouping methods with predefined parameters are used, then the inspection process can be automated, but the clustering process becomes time-consuming and lacks user control
Solution Approach 1:
The patent implements dynamic parameter adjustment during the clustering process, allowing users to modify clustering parameters in real-time based on intermediate results. This enables the system to adapt the clustering depth and criteria dynamically, reducing unnecessary computation time while maintaining automation.
Solution Approach 2:
The system incorporates feedback mechanisms where clustering results are evaluated and used to adjust subsequent clustering operations. This feedback loop allows the system to learn from previous clustering outcomes and optimize the process, reducing time consumption through intelligent parameter adjustment rather than fixed predefined parameters.
2Extent of automation
If traditional pattern grouping methods with predefined parameters are used, then the inspection process can be automated, but user control over the clustering process is lost
Solution Approach 1:
The system provides dynamic parameter adjustment capabilities that allow users to modify clustering parameters during the inspection process. This maintains automation while giving users control over the clustering behavior when needed, resolving the contradiction between automation and user control.
Solution Approach 2:
The system implements a hybrid approach where routine clustering operations run automatically, but users can intervene and adjust parameters when desired. This self-service model allows the system to handle standard cases autonomously while providing user control for specialized situations.
3Measurement precision
If massive data from SEM images is analyzed, then inspection accuracy improves, but the complexity and time consumption of pattern grouping increases
Solution Approach 1:
The patent divides the massive SEM image data into smaller pattern units and processes them through hierarchical clustering. This segmentation reduces the complexity of handling entire images at once while maintaining inspection accuracy by analyzing individual patterns and their relationships systematically.
Solution Approach 2:
The system dynamically adjusts the clustering depth and granularity based on the data characteristics and inspection requirements. This dynamic approach optimizes the balance between processing complexity and inspection accuracy, avoiding unnecessary computational overhead while maintaining high precision.
4Measurement precision
If massive data from SEM images is analyzed, then inspection accuracy improves, but processing time increases
Solution Approach 1:
By segmenting SEM images into discrete patterns and applying hierarchical clustering, the system processes data in manageable units. This approach maintains high inspection accuracy through detailed pattern analysis while reducing overall processing time compared to analyzing entire images sequentially.
Solution Approach 2:
The system performs preliminary pattern extraction and Fourier transform operations before the main clustering process. This preliminary processing organizes the data in advance, making the subsequent accuracy-critical analysis faster and more efficient.
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
This approach enables more efficient and accurate inspection by grouping similar patterns, allowing for improved hotspot analysis and wafer inspection, reducing the time and resource consumption in identifying defects and optimizing the manufacturing process.
Implementation Method 1
separating the plurality of patterns after Fourier Transform into multiple sets of patterns
Data Source
AI summary
Apparatuses, systems, and methods for grouping a plurality of patterns extracted from image data are disclosed. In some embodiments, the method for grouping the patterns comprises receiving the image data including the plurality of patterns that represent features to be formed on a portion of a wafer. The method also comprises separating the plurality of patterns after Fourier Transform into multiple sets of patterns. The method further comprises performing, to a respective set of patterns, a hierarchical clustering to obtain a plurality of subsets of patterns by recursively evaluating features related to similarity between patterns within the respective set of patterns.