Wafer Inspection Sensitivity via Local Attribute Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current inspection systems for semiconductor wafers face challenges in determining appropriate sensitivity parameters for defect detection, leading to inefficient defect detection due to inappropriate sensitivity settings, especially as feature sizes on wafers decrease, making it difficult to differentiate between defects and noise.
Innovation Solution
A method that determines local attributes, such as critical radius, for different locations on a wafer based on potential fault mechanisms and uses these attributes to adjust sensitivity settings dynamically, allowing for tailored defect detection across the wafer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single sensitivity parameter is used for the entire wafer, then the inspection process is simple to implement, but defect detection accuracy deteriorates because different regions have different local attributes requiring different sensitivity levels
Solution Approach 1:
The patent applies local quality by determining sensitivity parameters based on local attributes at different locations within the wafer design. Instead of using a uniform sensitivity parameter across the entire wafer, the system calculates location-specific sensitivity values that account for variations in local geometric attributes, pattern density, and defect criticality. This allows each region to be inspected with the appropriate sensitivity level, improving overall defect detection accuracy while maintaining a systematic approach to parameter determination.
Solution Approach 2:
The patent segments the wafer design into multiple locations or regions, each with its own locally determined sensitivity parameter. By dividing the inspection process into location-specific segments rather than applying a single global parameter, the system can tailor sensitivity settings to match the specific characteristics of each region, thereby resolving the contradiction between measurement precision and device complexity.
2Measurement precision
If sensitivity parameters are optimized for each location based on local attributes, then defect detection accuracy improves, but the computational complexity and time required for determining sensitivity parameters increases
Solution Approach 1:
The patent performs preliminary action by determining sensitivity parameters based on local attributes before the actual inspection process begins. By pre-calculating sensitivity values for different locations using design data and local geometric attributes, the system avoids the need for time-consuming real-time calculations during inspection. This preliminary determination of sensitivity parameters maintains high defect detection accuracy while significantly reducing the time loss during the actual inspection operation.
3Measurement precision
If higher sensitivity is used to detect smaller defects, then defect detection capability improves, but false positives from noise increase
Solution Approach 1:
The patent applies local quality by setting sensitivity parameters that are specific to each location's characteristics rather than using a uniformly high sensitivity across the entire wafer. By determining sensitivity based on local attributes such as pattern density, feature size, and defect criticality at each location, the system achieves high defect detection capability where needed while maintaining lower sensitivity in regions where noise is more prevalent, thereby reducing false positives without compromising overall detection capability.
Data Source
AI summary
Methods and systems for generating an inspection process for a wafer are provided. One computer-implemented method includes separately determining a value of a local attribute for different locations within a design for a wafer based on a defect that can cause at least one type of fault mechanism at the different locations. The method also includes determining a sensitivity with which defects will be reported for different locations on the wafer corresponding to the different locations within the design based on the value of the local attribute. In addition, the method includes generating an inspection process for the wafer based on the determined sensitivity. Groups may be generated based on the value of the local attribute thereby assigning pixels that will have at least similar noise statistics to the same group, which can be important for defect detection algorithms. Better segmentation may lead to better noise statistics estimation.