SiC Substrate Defect Classification via Reflected and Photoluminescence Light
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current defect inspection methods for silicon carbide substrates struggle to accurately distinguish basal plane dislocations from other defects, such as carrot defects, due to similar line-shaped images in photoluminescence light inspections, which hampers the production yield of high-voltage semiconductor devices.
Innovation Solution
A defect classifying method and apparatus that uses both reflected light and photoluminescence light inspections in parallel, employing ultraviolet illumination and band-edge emission wavelengths to differentiate basal plane dislocations, carrot defects, and other crystal defects by analyzing the shape and presence of defect images, enabling precise classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If photoluminescence light inspection is used to detect defects on silicon carbide substrate, then crystal defects can be detected, but basal plane dislocations cannot be distinguished from other defects such as carrot defects due to similar line-shaped images
Solution Approach 1:
The inspection method segments the defect analysis into two distinct observation modes: photoluminescence light inspection for detecting crystal defects, and reflected light inspection for observing surface morphology. By dividing the inspection into these two segments, the patent enables differentiation between basal plane dislocations and carrot defects, which appear similar in photoluminescence images alone.
Solution Approach 2:
The patent adds another dimension to defect inspection by introducing reflected light observation as a complementary method to photoluminescence inspection. While photoluminescence provides information about crystal defects in the bulk material, reflected light adds surface morphology information, creating a two-dimensional characterization space that enables accurate defect classification.
2Device complexity
If only photoluminescence light inspection is used, then inspection process is simple, but defect classification accuracy is insufficient
Solution Approach 1:
The patent merges two inspection methods (photoluminescence light inspection and reflected light inspection) into a unified defect classification system. The classification unit integrates information from both inspection results to accurately distinguish between different defect types, achieving high classification accuracy while maintaining operational simplicity through automated image processing.
Solution Approach 2:
The patent introduces an image processing unit and classification unit as intermediaries that automatically process and interpret the combined inspection data. These intermediary components bridge the gap between the two inspection methods and the final defect classification, enabling accurate results without requiring complex manual analysis.
3Measurement precision
If reflected light inspection is used to detect surface defects, then surface morphology can be observed, but crystal defects inside the substrate cannot be detected
Solution Approach 1:
The inspection system segments the observation into two functional parts: reflected light inspection for surface morphology and photoluminescence inspection for internal crystal defects. This segmentation allows each method to perform its specialized function while the classification unit integrates both information streams for comprehensive defect analysis.
Solution Approach 2:
The unified inspection apparatus performs multiple functions through a single integrated system: it detects surface defects via reflected light, detects internal crystal defects via photoluminescence, and classifies all defect types by combining both observation results. This multi-functionality eliminates the need for separate inspection equipment for surface and bulk defects.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for accurate detection and classification of basal plane dislocations and carrot defects, improving the production yield of high-voltage semiconductor devices by enhancing defect classification accuracy and distinguishing them from other crystal defects.
Implementation Method 1
projecting illumination light of an ultraviolet region toward the silicon carbide substrate... photoluminescence light emitted from the silicon carbide substrate
Implementation Method 2
separating, from the condensed light, each of the reflected light and the photoluminescence light... detecting each of the separated reflected light and the separated photoluminescence light
Data Source
AI summary
Provided are a defect classifying method and an inspection apparatus which are capable of classifying a defect by distinguishing a basal plane dislocation, which is a killer defect in bipolar high-voltage elements, from other defects. The defect classifying method according to the present invention includes: projecting an illumination beam toward a silicon carbide substrate and forming a reflection image and a photoluminescence image; a first inspection step of detecting a defect image from the reflection image formed; a second inspection step of detecting a defect image from the photoluminescence image formed; and a defect classification step of classifying detected defects based on whether or not the defect image is detected and the shape of the detected defect image.