Inspection Recipe Generation Using Design Context Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for creating inspection recipes in semiconductor manufacturing are inefficient, as they do not utilize design data, leading to manual, time-consuming processes with high variability and a high rate of nuisance defect detection, which reduces sensitivity and increases costs.
Innovation Solution
A computer-implemented method that creates an inspection recipe by mapping design attributes to image characteristics from a first design, allowing for the automatic generation of an inspection recipe for a second design using the same manufacturing process, thereby leveraging learned characteristics to optimize defect detection and reduce nuisance data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manual methods are used to create inspection recipes, then flexibility in handling different designs is maintained, but the process becomes time-consuming and labor-intensive
Solution Approach 1:
The system performs preliminary actions by pre-processing design data to extract geometric features, create context maps, and pre-determine inspection parameters before actual inspection begins. This preparation work is done in advance using automated algorithms that analyze the design layout and pre-calculate optimal inspection settings, thereby reducing the time required during the actual recipe creation phase while maintaining adaptability to different designs.
Solution Approach 2:
The system creates simplified representations or copies of the design data in the form of context maps that capture essential geometric features without requiring the full complexity of the original design files. These context map copies can be quickly generated and processed, enabling rapid recipe creation while preserving the necessary information for accurate inspection parameter determination across different designs.
2Device complexity
If uniform inspection sensitivity is applied across all design regions, then the inspection process is simplified, but nuisance defect detection increases and sensitivity to critical defects decreases
Solution Approach 1:
The system applies local quality by determining different inspection parameters for different regions of the design based on their specific geometric characteristics. The context map divides the design into multiple regions, and the system automatically selects appropriate inspection sensitivity levels, optics modes, and detection algorithms for each region according to its local features, thereby optimizing defect detection without uniformly complicating the entire inspection process.
Solution Approach 2:
The inspection process is segmented into multiple independent region-based inspection tasks. The context map divides the design layout into distinct regions with similar geometric characteristics, allowing the system to apply tailored inspection parameters to each segment. This segmentation enables the system to maintain relatively simple overall process control while achieving high measurement precision through region-specific optimization.
3Adaptability or versatility
If design data is not utilized in recipe creation, then the inspection system operates independently of design specifics, but the rate of nuisance defect detection increases
Solution Approach 1:
The system introduces context maps as an intermediary between the design data and the inspection process. The context map serves as a mediator that translates design geometric features into inspection-relevant information, allowing the inspection system to utilize design data effectively without requiring direct access to complex design files. This intermediary structure enables the system to reduce nuisance defect detection by understanding design-specific features while maintaining operational independence.
Solution Approach 2:
The system performs preliminary analysis of design data to create context maps that identify regions prone to nuisance defects based on their geometric characteristics. By pre-processing the design data to extract and flag potential nuisance-prone areas, the system can adjust inspection parameters in advance to minimize false detections, thereby reducing the nuisance defect rate while still operating independently during the actual inspection process.
Data Source
AI summary
Systems and methods for creating inspection recipes are provided. One computer-implemented method for creating an inspection recipe includes acquiring a first design and one or more characteristics of output of an inspection system for a wafer on which the first design is printed using a manufacturing process. The method also includes creating an inspection recipe for a second design using the first design and the one or more characteristics of the output acquired for the wafer on which the first design is printed. The first and second designs are different. The inspection recipe will be used for inspecting wafers after the second design is printed on the wafers using the manufacturing process.