Wafer Singulation Edge Discrepancy Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current singulation processes for electronic and integrated circuit devices require complex image processing techniques to characterize and control the separation of devices from substrates, necessitating a need for simplified methods to extract information from images for effective process monitoring and control.
Innovation Solution
A machine vision system captures images of singulated IC devices to identify discrepancies in peripheral edges, such as roughness, chips, and misalignment, allowing for real-time assessment and adjustment of singulation process variables like feed rates, temperatures, and alignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine vision techniques are applied to characterize singulation operations, then measurement precision is improved, but device complexity increases due to complicated image processing techniques required
Solution Approach 1:
The patent extracts and compares only the critical edge segments from the full device images, rather than processing entire images. This selective extraction of relevant features (edge segments) from the complete image data reduces processing complexity while maintaining measurement precision for singulation quality assessment
Solution Approach 2:
The patent divides the image processing task into segments by comparing edge segments rather than entire images. This segmentation approach breaks down the complex image processing into manageable portions (edge segment extraction, alignment, and comparison), reducing overall system complexity while preserving measurement accuracy
2Manufacturing precision
If comprehensive image processing is performed to identify all edge discrepancies, then manufacturing precision is improved, but productivity decreases due to processing time
Solution Approach 1:
The patent applies local quality by focusing processing resources only on the edge segments that are critical for singulation quality, rather than uniformly processing entire images. This localized approach to quality inspection maintains high manufacturing precision for critical features while improving productivity by reducing unnecessary processing of non-critical areas
Solution Approach 2:
The patent performs partial action by comparing only the necessary edge segments rather than conducting exhaustive processing of all image data. This partial processing approach provides sufficient quality control for singulation operations while significantly reducing processing time and improving productivity
Data Source
AI summary
A method for monitoring and controlling a substrate singulation process is described. Device edges are imaged and identified for analysis. Discrepancies in device edges are noted and used to modify a singulation process and to monitor the operation of singulation processes for anomalous behavior.