Wafer Cassette Mapping With Deep Learning Slot Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current substrate-mapping solutions in semiconductor and allied industries, such as those used in semiconductor fabs and solar-cell production, face challenges in accurately determining the classification state of substrates within carriers due to varying substrate diameters, types, and thicknesses, leading to issues like cross-loading, double-loading, and protrusions, which existing technologies struggle to address effectively.
Innovation Solution
A deep-convolutional neural-network-based substrate-mapping system that uses image acquisition and processing to classify the state of substrates within carriers, employing a pre-trained network to automatically tag each location as empty, properly loaded, double-loaded, cross-loaded, or protruded, and can handle diverse substrate types and thicknesses by capturing and analyzing images with minimal lighting requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional substrate-mapping solutions are used, then the system is simpler to implement, but the measurement precision and reliability of substrate state classification deteriorate due to varying substrate diameters, types, and thicknesses
Solution Approach 1:
The system changes the approach from traditional geometric and optical parameter-based detection to deep learning parameter extraction. The neural network learns optimal parameters automatically from training data, enabling accurate classification of substrates with varying diameters, types, and thicknesses without requiring complex manual parameter tuning
Solution Approach 2:
The patent replaces traditional mechanical and optical sensing systems with an image-based deep learning system. Instead of using complex mechanical sensors or optical interferometers, the system uses standard imaging combined with convolutional neural networks to achieve high-precision substrate state detection
2Productivity
If manual inspection methods are used, then the device complexity is lower, but the productivity and time efficiency deteriorate as inspection of entire carriers takes excessive time
Solution Approach 1:
The system performs preliminary action by pre-training the neural network with extensive substrate images before actual inspection. This preliminary training phase enables the network to rapidly classify substrates during production, achieving high-speed inspection without requiring complex real-time processing algorithms
Solution Approach 2:
The system creates digital copies (images) of substrates and processes them through the neural network. This copying approach allows parallel processing of multiple substrate images simultaneously, dramatically increasing inspection throughput compared to sequential manual or mechanical inspection methods
3Reliability
If existing mapping technologies are used, then the equipment setup is simpler, but the reliability of detecting cross-loading, double-loading, and protrusions deteriorates
Solution Approach 1:
The system transitions from two-dimensional image capture to three-dimensional substrate state classification. The neural network analyzes images to infer depth information and spatial relationships, enabling reliable detection of protrusions, cross-loading, and double-loading conditions without requiring complex multi-angle imaging or 3D scanning hardware
Data Source
AI summary
Various examples include a system and network to map of substrates within a substrate carrier (e.g., such as silicon wafers within a wafer cassette), and a classification of a state of each substrate, as well as the carrier in which the substrates are placed. In various examples provided herein, an image acquisition system, such as a camera, acquires multiple images of the substrates within the carrier. The image or images are then processed with a deep-convolutional neural-network to classify a state of the substrate relative to a substrate slot including empty slots, occupied slots (e.g., properly loaded slots), double-loaded slots, cross-slotted, and protruded (where a substrate is not fully loaded into a slot).