Semiconductor Defect Ratio Prediction via Foreign Substance Quantification

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for analyzing defects in semiconductor devices, such as the fail-bit analyzing method, struggle with low accuracy in quantifying characteristic defective ratios and foreign-substance defective ratios, leading to inadequate identification of defect causes and inefficient yield improvement.

Innovation Solution

A method that calculates actual defective ratios and critical areas for each fail bit mode, then uses these values to estimate the number of foreign substances in each process, allowing for the quantification of characteristic and foreign-substance defective ratios with high accuracy by minimizing gaps between predicted and actual defective ratios.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the fail-bit analyzing method is used to quantify defective ratios, then defect analysis can be performed, but the accuracy in quantifying characteristic defective ratios and foreign-substance defective ratios is low

Engineering Contradiction:

Improveaccuracy in quantifying defective ratiosVSAvoididentification of defect causes

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The patent segments the defective ratio into two distinct components: characteristic defective ratio (systematic defects) and foreign-substance defective ratio (random defects). By separating these previously mixed defect types, the method enables independent quantification of each component, thereby improving measurement precision and reliability in identifying specific defect causes.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces new calculation parameters including separate defective ratio expressions for characteristic and foreign-substance defects, along with specific calculation methods using fail bit mode ratios. These parameter changes transform the single defective ratio metric into multiple differentiated metrics, enabling accurate quantification of different defect types.

Inventive Principle:

Principle #35Parameter changes

2Productivity

If conventional defect analysis methods are used, then some defect information can be obtained, but the identification of defect sources is inadequate and yield improvement is inefficient

Engineering Contradiction:

Improveyield improvement efficiencyVSAvoiddefect cause identification

Core Design Contradiction:

ProductivityVSLoss of information

Solution Approach 1:

The patent establishes a feedback mechanism where calculated characteristic and foreign-substance defective ratios are used to identify specific defect causes, which then inform targeted process improvements. This closed-loop approach ensures that defect analysis results directly guide yield improvement actions, making the process efficient and data-driven.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent introduces calculated defective ratio values as intermediary metrics that bridge raw defect data and actionable insights. These intermediate calculations serve as mediators that translate complex defect patterns into quantifiable measures that directly indicate cause identification and guide yield improvement strategies.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS8612811B2Defective-ratio predicting method, defective-ratio predicting program, managing method for semiconductor manufacturing apparatus, and manufacturing method for semiconductor device

Publication Date: 2013.12.17 RENESAS ELECTRONICS CORP

AI summary

In a managing system for a semiconductor manufacturing apparatus, a predicting unit 121 predicts a characteristic defective ratio and a foreign-substance defective ratio of each process obtains an actual defective ratio of each fail bit mode and a critical area of each process and each fail bit mode, calculates the number of foreign substances of each process by using the actual defective ratio of each fail bit mode and the critical area of each process and each fail bit mode, the fail bit mode being except for an arbitrary fail bit mode, calculates a foreign-substance defective ratio of each process and a foreign-substance defective ratio of each fail bit mode by using the number of foreign substances, and calculates a characteristic defective ratio of the arbitrary fail bit mode based on the foreign-substance defective ratio and actual defective ratio of each fail bit mode.