Pattern Matching for Periodic Structures in SEM
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing pattern matching methods in scanning electron microscopes fail to accurately position periodic structures on semiconductor wafers, leading to unnecessary movement of the field of view and potential errors due to the selection of patterns based solely on coincidence, which can result in incorrect position checking and inefficient operation.
Innovation Solution
A method that determines the presence of a periodic structure in the sample using design data, allowing for pattern selection based on distance from the origin when a periodic structure is present and coincidence when it is not, thereby reducing unnecessary movement and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If pattern selection is based solely on coincidence with template, then matching speed is improved, but field of view movement becomes unnecessarily large causing positioning errors
Solution Approach 1:
The patent changes the selection parameter from purely coincidence-based to a hybrid approach that incorporates distance from the center point. When periodic structures are detected, the system prioritizes patterns closer to the center, thereby reducing unnecessary field of view movement while maintaining matching speed through the efficiency of the updated selection criterion.
Solution Approach 2:
The system incorporates feedback by detecting whether a periodic structure exists in the pattern before applying the selection rule. This feedback mechanism allows the system to adapt its behavior: using coincidence-based selection when appropriate and distance-based selection when periodic structures are present, thus optimizing both speed and precision dynamically.
2Measurement precision
If any pattern is selected from periodic structures, then field of view movement is minimized, but pattern selection becomes arbitrary reducing measurement reliability
Solution Approach 1:
The system uses feedback to detect the presence of periodic structures and adjusts its selection strategy accordingly. When periodic structures are detected, it applies distance-based selection to minimize field of view movement; when they are not present, it reverts to coincidence-based selection to ensure reliable pattern identification, thus maintaining measurement reliability through adaptive feedback control.
3Device complexity
If coincidence-based selection is used for periodic patterns, then matching process is simplified, but erroneous field of view movement increases
Solution Approach 1:
The patent introduces dynamics to the matching process by making the selection criterion flexible rather than fixed. The system dynamically switches between coincidence-based and distance-based selection strategies depending on whether periodic structures are detected, thereby maintaining simplicity when needed while improving precision when periodic patterns are present.
Solution Approach 2:
The system changes the selection parameter dynamically based on the detected pattern characteristics. By detecting periodic structures and switching to distance-based selection in those cases, the system reduces erroneous field of view movement without significantly increasing process complexity, as the additional detection step is performed only when necessary.
Data Source
AI summary
A pattern matching method which is capable of selecting a suitable measurement object pattern, even on a sample containing a periodic structure, and a computer program for making a computer execute the pattern matching. In a pattern matching method which executes matching between the design data-based first image of an object sample, and a second image, whether or not a periodic structure is included in a region to execute the matching is determined, so as to select a pattern, based on distance between an original point which is set in said image, and the pattern configuring said periodic structure, in the case where the periodic structure is included in said region, and to select a pattern based on coincidence of the pattern in said image, in the case where the periodic structure is not included in said region, and a computer program product.