Wafer Edge Feature Detection Using Third Derivative ZDDD Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional methods for characterizing semiconductor wafer topography, particularly in the edge and roll-off regions, are inadequate for detecting and quantifying features accurately, as they fail to effectively handle perturbations and anomalies, especially in negative ZDD regions.
Innovation Solution
The method involves calculating the first and third derivatives of the ZDD profile to obtain the ZDDD profile, applying bilateral filtering to suppress noise, and reconstructing feature profiles to enhance detection and quantification of wafer edge features, including the use of polar coordinates and adaptive baseline determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional thickness measurements and second derivative processing are used, then the measurement process is simple, but the detection accuracy of wafer edge features is insufficient
Solution Approach 1:
The patent segments the wafer surface into multiple measurement zones (center, mid-ring, edge, roll-off) and applies different analysis methods to each zone. This allows targeted detection of edge features using third derivative processing specifically in regions where it is most beneficial, rather than applying complex processing uniformly across the entire wafer surface.
Solution Approach 2:
The patent transitions from traditional second derivative (curvature) analysis to third derivative (rate of change of curvature) analysis, adding a new dimensional perspective to the measurement. This enables detection of subtle edge perturbations that are invisible to second derivative methods, particularly in the roll-off region where the wafer thickness naturally varies.
2Reliability
If wafer edge region measurements are performed, then comprehensive topography characterization is achieved, but measurement reliability is reduced due to slope roll off
Solution Approach 1:
The patent applies local quality by using polar coordinates centered at the wafer edge rather than at the wafer center. This coordinate transformation aligns the measurement system with the natural geometry of the wafer edge, making the roll-off region appear as a gradual, predictable variation rather than a sharp discontinuity. This improves measurement reliability in the edge region while maintaining the ability to detect local features.
Solution Approach 2:
The patent changes the mathematical parameters from Cartesian coordinates to polar coordinates, and from second derivative to third derivative processing. These parameter changes transform the measurement equations to better suit the geometry and characteristics of wafer edge regions, improving the reliability of feature detection in these challenging areas.
3Adaptability or versatility
If conventional feature detection methods are used, then the detection scope is limited, but the processing time is short
Solution Approach 1:
The patent performs preliminary action by pre-defining multiple feature detection modes (Type 1 bumps, Type 2 bumps, edge perturbations, roll-off variations) with pre-configured analysis parameters for each. During measurement, the system can quickly select and execute the appropriate pre-configured mode without requiring complex real-time decision-making, thus expanding detection scope while maintaining short processing times.
Data Source
AI summary
Disclosed herein is a method to enhance detection and quantification of features in the wafer edge/wafer roll off regions. Modifications and improvements have been made to earlier methods which enable improved accuracy and increased scope of feature detection.