Optical Waveform Analysis for Semiconductor Defect Localization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing inspection techniques for semiconductor devices struggle to effectively analyze defective parts, making it difficult to specify the location of defects in complex circuit elements.

Innovation Solution

An analysis method involving optical measurement that calculates the degree of correspondence between waveform data from multiple positions on a defective and non-defective semiconductor device, allowing for the identification of defective parts based on this correspondence.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If optical measurement is performed on semiconductor devices with many circuit elements, then measurement data is obtained, but it becomes difficult to specify defective parts effectively

Engineering Contradiction:
Improvedefect detection accuracyVSAvoidcomplexity of analysis
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the semiconductor device into multiple measurement regions and divides the analysis into two steps: first performing pattern recognition to identify candidate defective regions, then performing detailed waveform analysis only in those regions. This segmentation approach reduces the complexity of analyzing the entire device while maintaining defect detection accuracy.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts and removes noise components from the measurement waveforms through spectral analysis and filtering techniques. By separating the noise components from the actual defect signals, the analysis can focus on relevant information, reducing complexity while improving defect specification accuracy.

Inventive Principle:
Principle #2Taking out (Extraction)

2Measurement precision

If detailed waveform analysis is performed for all positions, then defect detection accuracy is improved, but analysis time increases

Engineering Contradiction:
Improvedefect specification accuracyVSAvoidanalysis time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary pattern recognition analysis on all measurement positions to identify candidate defective regions before conducting detailed waveform analysis. This preliminary action filters out non-defective regions, so that time-consuming detailed analysis is performed only where necessary, significantly reducing total analysis time while maintaining defect specification accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

Instead of performing exhaustive detailed waveform analysis on all positions, the patent applies partial analysis only to candidate defective regions identified by pattern recognition. This partial action approach achieves sufficient defect specification accuracy without the time cost of complete analysis across the entire device.

Inventive Principle:
Principle #16Partial or excessive action

3Measurement precision

If measurement data is collected from multiple positions, then defect localization capability is improved, but data processing complexity increases

Engineering Contradiction:
Improvedefect localization accuracyVSAvoiddata processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent creates a pattern recognition model based on characteristics of defective waveforms and uses this model to compare against measurement data from multiple positions. This copying approach allows systematic identification of defective regions without manually processing each waveform, reducing data processing complexity while maintaining multi-position defect localization capability.

Inventive Principle:
Principle #26Copying

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 enables effective analysis of defective semiconductor devices by identifying and localizing defects through the distribution of correspondence in time waveforms, enhancing the accuracy of defect detection.

Implementation Method 1

a light detector configured to detect light from the semiconductor device to which a test signal has been input

Methodology Applied
Scientific EffectLight detection: Photoelectric Effect

Data Source

PatentEP3719513B1Analysis method, analysis device, analysis program, and recording medium for recording analysis program
Publication Date: 2025.10.15 HAMAMATSU PHOTONICS KK
  • EP3719513B1 patent drawingFigure 1
  • EP3719513B1 patent drawingFigure 2
  • EP3719513B1 patent drawingFigure 3

AI summary

An inspection apparatus includes a light sensor that detects light from a semiconductor device to which an electric signal has been input, an optical system that guides light from the semiconductor device to the light sensor, and a control device electrically connected to the light sensor. The control device includes a measurement unit that acquires waveform data obtained by optical measurement for each of a plurality of positions on a defective semiconductor device and waveform data obtained by the optical measurement for each of a plurality of positions on a non-defective semiconductor device, a calculation unit that calculates a degree of correspondence between the waveform data of the defective semiconductor device and the waveform data of the non-defective semiconductor device, and an analysis unit that analyzes a defective part of the defective semiconductor device on the basis of the degree of correspondence for each of the plurality of positions.