Automated Recess Dimension Measurement in Semiconductor Imaging

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In semiconductor manufacturing, manual dimension measurement of recesses in semiconductor devices using scanning electron microscopes is time-consuming and prone to person-dependent errors, especially when measuring multiple recesses.

Innovation Solution

An information processing apparatus that automates the detection of recess regions, film boundaries, and contours in images, using frequency analysis, Fourier transforms, and change point detection algorithms to improve measurement efficiency and accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual dimension measurement is performed for multiple recesses, then measurement accuracy can be maintained, but measurement time increases significantly

Engineering Contradiction:

Improvedimension measurement accuracyVSAvoidmeasurement time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent uses image data as a copy of the physical recesses to perform measurements. By capturing images of the recesses and analyzing them computationally, the system avoids direct manual measurement of each physical recess, thereby reducing measurement time while maintaining accuracy through automated image analysis algorithms

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent replaces manual mechanical measurement operations with automated image processing and computational analysis. The measurement system uses software algorithms to detect recess boundaries and calculate dimensions from images, substituting the mechanical manual measurement process with an automated digital system that operates faster and more consistently

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If manual measurement operations are performed, then flexibility in measurement selection is maintained, but person-dependent errors occur

Engineering Contradiction:

Improvemeasurement selection flexibilityVSAvoidmeasurement consistency

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent implements automated feedback mechanisms where the system automatically detects recess regions, identifies measurement targets, and performs measurements based on predefined criteria. The measurement target selection unit uses feedback from image analysis to automatically determine which recesses to measure, eliminating person-dependent variability while maintaining measurement flexibility through programmable selection criteria

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The measurement system performs self-service by automatically selecting measurement targets and executing measurements without human intervention. The system autonomously processes images, identifies recesses, selects measurement candidates based on predefined rules, and carries out dimension measurements, thereby eliminating person-dependent errors while maintaining adaptability through configurable measurement parameters

Inventive Principle:

Principle #25Self-service

3Quantity of substance

If multiple recesses are measured manually, then comprehensive data collection is achieved, but measurement efficiency decreases

Engineering Contradiction:

Improvenumber of recesses measuredVSAvoidmeasurement efficiency

Core Design Contradiction:

Quantity of substanceVSProductivity

Solution Approach 1:

The patent enables continuous automated measurement of multiple recesses by processing images sequentially without interruption. The system continuously analyzes image data, detects recess regions, selects measurement targets, and calculates dimensions in an automated workflow that operates without the interruptions inherent in manual measurement, thereby increasing productivity while measuring the same number of recesses

Inventive Principle:

Principle #20Continuity of useful action

Solution Approach 2:

The patent segments the measurement process into distinct automated stages: image acquisition, recess region detection, measurement target selection, and dimension calculation. By dividing the overall measurement task into these manageable segments that can be processed automatically and in parallel where possible, the system increases efficiency while maintaining comprehensive data collection across multiple recesses

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20230360190A1Profile detection method, profile detection program, and information processing apparatus

Publication Date: 2023.11.09 TOKYO ELECTRON LTD

AI summary

A profile detection method includes a region detection step of detecting, by analyzing data of an image in which a plurality of recesses recessed in one direction are arranged in an intersecting direction with respect to the one direction, each recess region in the image, a boundary detection step of detecting a boundary of a film included in the image by analyzing the data, and a contour detection step of detecting a contour of the recess for the each recess region in the image by analyzing the data.