After-Development Image Comparison for Etch Defect Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current defect classification methods in semiconductor manufacturing do not accurately predict defects after etching, as they do not consider etch conditions, leading to inaccurate defect rates and reduced yield.
Innovation Solution
A method involving training a model to compare after development images (ADI) with after etch images (AEI) to predict defectiveness, allowing for adjustments in etch conditions to reduce defect likelihood.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current defect classification methods are used without considering etch conditions, then the process is simpler and faster, but the defect prediction accuracy deteriorates
Solution Approach 1:
The patent performs defect classification before the etching process by analyzing after-development images and predicting which features will fail during etching. This preliminary action allows the system to identify at-risk features in advance and adjust etch conditions accordingly, improving prediction accuracy without requiring complex real-time monitoring during etching.
Solution Approach 2:
The patent establishes a feedback loop where defect predictions from after-development images are used to adjust etch conditions. The system compares predicted defective features with actual etch outcomes and uses this feedback to refine the classification model, continuously improving accuracy while managing process complexity through iterative optimization.
2Productivity
If etch conditions are not optimized based on defect predictions, then the manufacturing process is simpler, but the yield deteriorates
Solution Approach 1:
The system performs preliminary defect classification before etching by analyzing after-development images. By identifying which features are likely to fail during etching in advance, the system can pre-adjust etch conditions to prevent defects, thereby improving yield without requiring complex real-time control during the etching process itself.
Solution Approach 2:
The patent modifies etch process parameters (such as etch time, gas flow rates, or plasma power) based on the predicted defectiveness of specific features. By changing these parameters dynamically according to the classification results, the system optimizes yield while managing complexity through targeted parameter adjustments rather than complete process redesign.
3Measurement precision
If comprehensive defect analysis is performed, then the defect detection accuracy is improved, but the inspection time increases
Solution Approach 1:
The patent performs defect analysis during the after-development imaging step, which occurs before etching. By classifying features as potentially defective at this early stage, the system eliminates the need for time-consuming post-etch inspection of all features, achieving high detection accuracy while reducing overall inspection time through early identification and targeted verification.
Solution Approach 2:
The system extracts and focuses analysis on only the features that are predicted to be defective based on after-development image characteristics. By taking out the time-consuming analysis of non-critical features and focusing resources only on predicted defect locations, the system maintains high detection accuracy while significantly reducing total inspection time.
Data Source
AI summary
Described herein is a method of training a model configured to predict whether a feature associated with an imaged substrate will be defective after etching of the imaged substrate and determining etch conditions based on the trained model. The method includes obtaining, via a metrology tool, (i) an after development image of the imaged substrate at a given location, the after development image including a plurality of features, and (ii) an after etch image of the imaged substrate at the given location; and training, using the after development image and the after etch image, the model configured to determine defectiveness of a given feature of the plurality of features in the after development image. In an embodiment, the determining of defectiveness is based on comparing the given feature in the after development image with a corresponding etch feature in the after etch image.