Defect Observation Device Using Average Brightness Reference

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In semiconductor manufacturing, the miniaturization of design patterns and complexity of processes lead to diversified defects, increasing the likelihood of erroneous defect detection, especially when periodic patterns are not recognizable in defect images, making it challenging to generate suitable reference images for high-precision defect detection.

Innovation Solution

Calculating the defect occupation rate and determining a threshold value to generate a reference image using the average brightness value of pixels in the inspected image, allowing for stable high-precision defect detection even when patterns are not visible.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a reference image is generated using periodicity of patterns from defect images, then defect detection precision is improved, but the method becomes inapplicable when periodic patterns are not recognizable

Engineering Contradiction:

Improvedefect detection precisionVSAvoidapplicability to non-periodic patterns

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The invention changes the fundamental parameter used for reference image generation from pattern periodicity to average brightness value. This allows the system to generate reference images for both periodic and non-periodic patterns by simply changing the generation method based on image characteristics, thereby improving adaptability while maintaining detection precision

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The invention introduces a dynamic selection mechanism that determines whether to use periodicity-based or average brightness-based reference image generation based on the actual image characteristics. This dynamic adaptation allows the system to handle diverse defect types and pattern conditions effectively

Inventive Principle:

Principle #15Dynamics

2Measurement precision

If supersensitive defect inspection is executed to detect minute defects, then defect detection capability is improved, but the number of false defects increases

Engineering Contradiction:

Improvedefect detection capabilityVSAvoidfalse defect detection

Core Design Contradiction:

Measurement precisionVSObject-generated harmful factors

Solution Approach 1:

The invention uses the defect occupation rate as feedback to dynamically adjust the reference image generation method. When false defects are detected, the system can recalculate the defect occupation rate and switch to average brightness-based generation, thereby reducing false positives while maintaining sensitivity to real defects

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The invention replaces the mechanical pattern-matching approach with a statistical average brightness approach for reference image generation. This substitution reduces the system's sensitivity to pattern variations that cause false defects while maintaining the ability to detect actual defects through brightness anomalies

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Manufacturing precision

If defect coordinates are corrected from low-magnification images to acquire high-definition images, then image quality is improved, but processing time increases

Engineering Contradiction:

Improveimage qualityVSAvoidprocessing time

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The invention performs preliminary generation of reference images using average brightness values at low magnification before high-magnification imaging. This preliminary action allows for early defect coordinate identification and correction, reducing the overall processing time while maintaining high-definition image quality for final observation

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS9569836B2Defect observation method and defect observation device

Publication Date: 2017.02.14 HITACHI HIGH TECH CORP

AI summary

Cases in which defects are analyzed in a manufacturing process stage in which a pattern is not formed or in a manufacturing process in which a pattern formed on a lower layer does not appear in the captured image are increasing. However, in these cases, there is a problem of not being able to synthesize a favorable reference image and failing to detect a defect when a periodic pattern cannot be recognized in the pattern. In the present invention, a defect occupation rate, which is the percentage of an image being inspected occupied by a defect region, is found, it is determined whether the defect occupation rate is higher or lower than a threshold, and, in accordance with the determination results, it is determined whether to create, as the reference image, an image comprising pixels having the average luminance value of the luminance values of a plurality of pixels contained in the image being inspected. In particular, when the defect occupation rate is low, an image comprising pixels having the average luminance value of the luminance values of a plurality of pixels contained in the image being inspected is used as the reference image.