Semiconductor Defect Inspection Reference Region Selection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional defect inspection methods for semiconductor wafers face challenges in setting appropriate reference images due to variations in local brightness and surface roughness, leading to erroneous defect detection or overlooking of defects.

Innovation Solution

An inspection information generation device and method that acquires design information, extracts candidate regions, captures images, calculates similarity, and determines reference regions based on similarity or distance, using partial images with edge patterns to improve defect detection accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If design information is used to set reference regions with similar layout, then the reference image selection is simplified, but captured images may not be similar due to optical effects like scattered light from lower layer or peripheral patterns

Engineering Contradiction:

Improvereference image selection processVSAvoidimage similarity

Core Design Contradiction:

Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent replaces the conventional approach of using design information (layout-based selection) with an image-based similarity calculation approach. The system captures actual images of candidate regions and computes similarity metrics directly from the captured image data, substituting the theoretical layout-matching mechanism with empirical image-comparison mechanisms that account for actual optical effects and variations.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Device complexity

If conventional die comparison or cell comparison schemes are used, then processing is simplified, but scattered light from lower layer or peripheral patterns causes erroneous information or missed defects

Engineering Contradiction:

Improveinspection methodVSAvoiddefect detection accuracy

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The patent introduces dynamic adaptability by calculating similarity for multiple candidate regions and selecting the best match based on actual image data rather than fixed comparison schemes. The system dynamically adjusts the reference region selection based on the specific characteristics of each inspected region, allowing the inspection method to adapt to variations in scattered light patterns and optical conditions for different locations on the wafer.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent changes the selection criterion from layout similarity (design information) to image similarity (captured image data). By using similarity calculation based on actual captured images, the system changes the fundamental parameter used for reference region selection, enabling the method to account for optical effects and variations that are not visible in design information alone.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11041815B2Inspection information generation device, inspection information generation method, and defect inspection device

Publication Date: 2021.06.22 HITACHI HIGH TECH CORP

AI summary

An inspection information generation device includes a design information acquirer configured to acquire design information of a sample to be inspected, a candidate region extractor configured to use the design information to extract multiple candidate regions, an image capturer configured to capture images of the multiple candidate regions, a similarity calculator configured to use the images of the multiple candidate regions to calculate a similarity or distance between the multiple candidate regions, and a region determiner configured to use the similarity or the distance to determine, as inspection information, at least one reference region corresponding to a region to be inspected.