Semiconductor Wafer Defect Root Cause Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional inspection methods for semiconductor manufacturing defects, such as those caused by Cross SADP, cannot accurately identify the root cause of unexpected shifts in hole patterns, limiting the ability to address and correct manufacturing process issues.
Innovation Solution
A system and method that captures images of semiconductor wafers, calculates geometric centers and average centers, and performs coordinate transformations to determine the precise shift amount and identify the manufacturing process responsible for defects, enabling precise defect identification and process correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional optical inspection equipment is used to check hole patterns, then the inspection can be performed, but the root cause of unexpected shifts cannot be distinguished
Solution Approach 1:
The patent segments the inspection process into multiple stages: capturing multiple images at different magnifications, calculating geometric centers for each image, computing average geometric centers across images, and determining shifts by comparing these averages. This segmentation allows the system to distinguish between measurement noise and actual defects, thereby identifying the root cause of shifts while maintaining precision.
Solution Approach 2:
The patent introduces a new dimension of analysis by capturing multiple images at different magnifications and using coordinate transformations to relate features across images. This multi-dimensional approach enables the system to differentiate between apparent shifts and actual manufacturing defects, providing both precise shift measurement and root cause identification.
2Loss of information
If multiple images are captured and processed to identify shift causes, then root cause analysis is enabled, but the inspection process complexity increases
Solution Approach 1:
The patent creates a multi-functional inspection system that performs multiple tasks using the same hardware: capturing images at different magnifications, calculating geometric centers, determining shifts, and identifying root causes. By making the inspection apparatus universal, the patent avoids adding separate equipment for each function, thereby managing complexity while enabling comprehensive analysis.
Solution Approach 2:
The patent uses multiple copies of the same feature (geometric features appearing in multiple images at different magnifications) to verify measurements and identify defects. This copying approach allows the system to cross-validate results and distinguish between measurement errors and actual defects without requiring complex additional hardware.
3Ease of operation
If conventional single-image inspection is used, then the inspection process is simple, but the ability to provide precise shift amounts and locate specific defects is limited
Solution Approach 1:
The patent performs preliminary actions by capturing multiple images at different magnifications before final analysis. These preliminary images are used to calculate geometric centers and determine average positions, which then serve as the basis for precise shift measurements. This preliminary data collection enables high precision while maintaining operational simplicity through automated processing.
Data Source
AI summary
A system, method, and non-transitory computer-readable medium for identifying a cause of manufacturing defects are provided. The system includes a processing unit and an image capture unit electrically coupled to the processing unit. The system is configured to capture, by the image capture unit, a number N of images covering different portions of a semiconductor wafer, wherein each of the number N of images comprises a number M of geometric features. The system is further configured to specify a number M of serial numbers, each associated with one of the number M of geometric features. The system is further configured to calculate, by the processing unit, a geometric center for each of the geometric features of the number N of images. The system is further configured to calculate, based on the number N of images, a number M of average geometric centers associated with the number M of serial numbers. The system is further configured to calculate a shift amount for each geometric feature of the number N of images.