Fan-Out WLP Overlay Control Using Tool-Signal Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In fan-out wafer level packaging (WLP), managing vertical alignment or overlay errors between interconnection features is challenging due to variations in tool output signals from degraded alignment tools, leading to misalignment of features like through-InFO vias (TIVs) with corresponding connection features.
Innovation Solution
The use of machine learning techniques to analyze tool signals and correlate them with overlay metrology values, allowing for the prediction of overlay metrology and correction of overlay errors by adjusting tool settings or parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional overlay metrology processes are used to monitor vertical alignments, then alignment accuracy can be maintained, but tool degradation and variations in tool output signals lead to overlay errors
Solution Approach 1:
The patent implements a feedback mechanism by continuously monitoring tool output signals and using machine learning to correlate signal variations with overlay errors. The system measures actual overlay errors, compares them with predicted errors from tool signals, and uses this feedback to refine the machine learning model and adjust tool parameters, creating a closed-loop control system that compensates for tool degradation over time
Solution Approach 2:
The patent replaces traditional mechanical/optical overlay measurement systems with a machine learning-based predictive system. Instead of relying solely on physical measurement tools, the system uses computational models that analyze tool output signals and predict overlay errors, substituting direct measurement with intelligent prediction and correction
2Manufacturing precision
If machine learning techniques are used to predict overlay metrology, then overlay errors can be corrected, but system complexity increases
Solution Approach 1:
The patent segments the overlay control system into distinct functional modules: tool signal acquisition, machine learning model training, overlay error prediction, and correction implementation. This segmentation allows each module to be independently optimized, maintained, and scaled, reducing overall system complexity while maintaining high prediction accuracy
Solution Approach 2:
The patent introduces machine learning models as intermediary components between tool signals and overlay control decisions. These models act as intelligent mediators that translate raw tool signals into predictive overlay error information, bridging the gap between tool operation and quality control without requiring direct complex interactions
Data Source
AI summary
The current disclosure describes techniques for managing vertical alignment or overlay in semiconductor manufacturing using machine learning. Alignments of interconnection features in a fan-out WLP process are evaluated and managed through the disclosed techniques. Big data and machine learning are used to train a classification that correlates the overlay error source factors with overlay metrology categories. The overlay error source factors include tool signals. The trained classification includes a base classification and a Meta classification.