Fan-Out Wafer Defect Detection via Image-Based Alignment Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Embedded Wafer Level Packaging (eWLP) and Embedded wafer Ball Grade array (eWLB) processes face alignment errors and height differences between silicon dies and frames, leading to manufacturing yield loss and reliability issues due to assumptions of perfect alignment, which are not accounted for in equipment designed for full wafers.
Innovation Solution
A method for monitoring and correcting defects in the fan-out wafer manufacturing process through image acquisition and processing, allowing for corrective operations such as repositioning or updating manufacturing entities with misalignment data, to address issues like die misalignment, frame misalignment, and redistribution layer defects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If manufacturing equipment designed for full wafers is used for fan-out wafers, then device complexity is reduced, but manufacturing precision deteriorates due to unaccounted die location variations
Solution Approach 1:
The system performs preliminary actions by capturing images of the fan-out wafer at multiple manufacturing stages and computing die location information before subsequent manufacturing steps. This advance preparation of location data allows the system to compensate for die misalignment, rotation, and tilt variations before they affect manufacturing precision, thereby resolving the contradiction between using simpler equipment and maintaining high precision.
Solution Approach 2:
The system implements feedback by continuously monitoring die location through image processing at each manufacturing stage and using this information to adjust subsequent manufacturing operations. The computed die location data feeds back to correct for alignment errors, ensuring that manufacturing precision is maintained even when using equipment designed for full wafers rather than diced wafers.
2Ease of manufacture
If perfect alignment is assumed during manufacturing, then manufacturing process is simplified, but manufacturing precision deteriorates due to uncorrected die misalignment
Solution Approach 1:
The system performs preliminary measurement and computation of die location information before manufacturing steps that require precise alignment. By capturing images and computing location data in advance, the system prepares correction information without complicating the actual manufacturing process, thus maintaining ease of manufacture while improving precision.
Solution Approach 2:
The system uses feedback from image processing to provide correction information that compensates for die misalignment. This feedback mechanism allows the manufacturing process to remain simple while automatically adjusting for alignment errors through computational correction, resolving the contradiction between process simplicity and alignment precision.
3Productivity
If multiple manufacturing stages are performed without intermediate monitoring, then productivity is improved, but manufacturing precision deteriorates due to undetected defects
Solution Approach 1:
The system performs preliminary image capture and defect detection at intermediate stages between manufacturing steps. By identifying defects early through intermediate monitoring, the system can correct issues before they propagate through subsequent manufacturing stages, maintaining both productivity and precision.
Solution Approach 2:
The system implements feedback through intermediate monitoring that detects defects and provides information for corrective actions. This feedback loop allows the system to maintain high precision by catching and correcting defects early, while preserving productivity by using efficient image processing and targeted corrective operations rather than complete process rework.
Data Source
AI summary
A system and method for monitoring a manufacturing process of a fan-out wafer, the method may include acquiring a first set of images of dies after a completion of a first manufacturing stage of a manufacturing process of the fan-out wafer; processing the first set of images to detect defects; performing at least one corrective operation in response to at least one defect detected by processing the first set of images; acquiring a second set of images of dies after a completion of a second manufacturing stage of the manufacturing process of the fan-out wafer, the second manufacturing process follows the first manufacturing process; processing the second set of images to detect defects; and performing at least one corrective operation in response to at least one defect detected by processing the second set of images.