Defect Inspection Using Region-Specific Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional defect inspection methods for semiconductor substrates are inefficient due to low accuracy in classifying defects, requiring extensive manual inspection and increasing manufacturing time and cost, as they rely on uniform parameters for all regions despite varying background signals and defect types.
Innovation Solution
The method involves illuminating specific points on a semiconductor substrate with light and analyzing it using region-specific characteristic parameters such as contrast, intensity, and brightness to accurately detect and classify defects without the need for subsequent manual inspection, employing a scanning unit, determining unit, and inspecting unit to differentiate between inspection regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If uniform parameters are used for inspecting all regions of the semiconductor substrate, then the inspection process is simple and fast, but the defect classification accuracy is low requiring manual review
Solution Approach 1:
The patent applies local quality by dividing the semiconductor substrate into multiple inspection regions and assigning different characteristic parameters to each region based on its specific background signal characteristics. This allows each region to be inspected with optimized parameters tailored to its local features, improving defect classification accuracy without sacrificing inspection speed.
Solution Approach 2:
The patent implements parameter changes by dynamically selecting different characteristic parameters (such as contrast, intensity, brightness) for different inspection regions based on their background signal characteristics. This enables the inspection system to adapt parameters to local conditions, achieving high accuracy automated classification across diverse regions.
2Measurement precision
If multiple inspection regions are set on a single semiconductor substrate to accommodate increased integration, then more defects can be detected, but the inspection time increases significantly
Solution Approach 1:
The patent applies segmentation by dividing the semiconductor substrate into multiple inspection regions, each with its own characteristic parameters. This allows parallel processing of different regions with optimized parameters, enabling comprehensive defect detection across the entire substrate while maintaining efficient inspection throughput through automated classification.
Solution Approach 2:
The patent implements self-service by enabling the inspection system to automatically classify defects using region-specific characteristic parameters without requiring manual intervention. The system autonomously determines defect types and makes pass/fail decisions, eliminating time-consuming manual review processes even when inspecting hundreds of regions.
3Measurement precision
If manual inspection is performed to accurately classify defects, then defect classification accuracy is high, but the reviewing time and manufacturing cost increase
Solution Approach 1:
The patent implements self-service by enabling the automated inspection system to classify defects autonomously using characteristic parameters specific to each inspection region. The system automatically determines defect types and makes classification decisions without human intervention, achieving both high accuracy and efficiency by eliminating manual review processes.
Solution Approach 2:
The patent replaces the mechanical manual inspection process with an automated optical inspection system that uses characteristic parameters for defect classification. This substitution eliminates the need for human reviewers while maintaining or improving classification accuracy through consistent, objective parameter-based analysis across all inspection regions.
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 significantly reduces reviewing time by accurately classifying defects, allowing for automated inspection and reducing manufacturing costs by eliminating the need for manual review processes.
Implementation Method 1
a light emitter for emitting light onto the semiconductor substrate
Implementation Method 2
to collect scattered or reflected light from the semiconductor substrate
Implementation Method 3
to collect scattered or reflected light from the semiconductor substrate
Data Source
AI summary
In a method of inspecting defects, a first actual region of an actual object is inspected based on a first characteristic parameter as an inspection condition. A point where an inspection region of the actual object is changed into a second actual region from the first actual region is determined. The second actual region is then inspected based on a second characteristic parameter as the inspection condition. The first and second parameters may include contrast of a light that is reflected from a reference object, intensity of the light, brightness of the light, a size of a minute structure on the reference object, etc. The characteristic parameters of each reference region on the reference object are set. Thus, the defects may be accurately classified so that a time and a cost for reviewing the defects may be markedly reduced.