Multi-Zone Alignment Modeling for Semiconductor Wafer Metrology
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing metrology tools face challenges in accurately determining and modeling different alignment zones on semiconductor wafers, which can lead to reduced throughput due to the need for user input and increased alignment residuals when using a single alignment model across the entire wafer.
Innovation Solution
A semiconductor tool and method that measures alignment data at multiple locations, selects an analysis area, divides it into zones with different alignment signatures, and models each zone using distinct alignment models to minimize alignment signal metric differences, thereby reducing alignment residuals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple zones are determined with user input to improve alignment modeling accuracy, then alignment precision is improved, but throughput decreases
Solution Approach 1:
The system automatically determines alignment zones and selects appropriate alignment models without requiring user input. The controller autonomously analyzes alignment data, identifies zones with different alignment signatures, and applies corresponding models to each zone, enabling the system to serve itself in the zone determination process while maintaining high alignment precision and throughput
Solution Approach 2:
The wafer is divided into multiple alignment zones based on detected alignment signatures. Each zone is independently analyzed and modeled with zone-specific alignment models, allowing precise local alignment correction while maintaining overall processing efficiency through automated zone identification
2Ease of operation
If a single alignment model is used across the entire wafer to simplify the process, then ease of operation is improved, but alignment precision deteriorates due to increased alignment residuals
Solution Approach 1:
Different alignment models are applied to different zones of the wafer based on locally detected alignment signatures. Each zone receives a customized alignment model that matches its specific alignment characteristics, ensuring high alignment precision locally while the automated system maintains overall operational simplicity
Solution Approach 2:
The system dynamically selects and applies different alignment models to different zones based on real-time analysis of alignment data. The controller automatically adapts the alignment modeling approach to match the specific alignment signature of each zone, optimizing alignment precision without requiring manual intervention
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 allows for more accurate alignment modeling across the wafer by automatically determining and modeling distinct zones, improving alignment precision and reducing residuals compared to using a single model.
Implementation Method 1
an illumination source to generate an illumination beam
Implementation Method 2
one or more illumination optical elements to direct a portion of the illumination beam to a sample
Implementation Method 3
one or more collection optical elements to direct radiation emanating from the sample to the detector
Data Source
AI summary
A semiconductor tool includes an illumination source to generate an illumination beam, one or more illumination optical elements to direct a portion of the illumination beam to a sample, a detector, one or more collection optical elements to direct radiation emanating from the sample to the detector, and a controller communicatively coupled to the detector. The controller is configured to measure alignment at a plurality of locations across the sample to generate alignment data, select an analysis area for alignment zone determination, divide the analysis area into two or more alignment zones having different alignment signatures; model the alignment data of at least a first alignment zone of the two or more alignment zones using a first alignment model, and model the alignment data of at least a second alignment zone of the two or more alignment zones using a second alignment model different than the first alignment model.