Semiconductor Defect Analysis via Current Data Segmentation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Semiconductor devices face defects due to time-dependent dielectric breakdown (TDDB) caused by continuous low voltage application, leading to leakage currents and failure to turn off normally, which are difficult to detect before shipping.

Innovation Solution

A method of analyzing defects in semiconductor devices by collecting current data, dividing it into components using specific functions for Fowler-Nordheim and trap-assisted tunneling, calculating quality indexes, and redistributing time data based on Weibull distribution to identify defect types and improve reliability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If continuous low voltage is applied to test semiconductor devices, then TDDB defects can be detected, but the testing process takes excessive time and reduces productivity

Engineering Contradiction:

Improvedefect detection accuracyVSAvoidtesting speed

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent segments the current data into multiple components (first component value and second component value) based on different voltage ranges. By dividing the current characteristics into segments, the method can analyze different aspects of TDDB defects separately, improving detection accuracy without requiring excessively long testing times. The first component captures early-stage defects while the second component captures later-stage defects, enabling comprehensive analysis through segmented measurement.

Inventive Principle:

Principle #1Segmentation

2Measurement precision

If detailed current data analysis is performed to accurately detect TDDB defects, then measurement precision improves, but the complexity of the analysis process increases

Engineering Contradiction:

Improvedefect detection precisionVSAvoidanalysis process complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The complex current data is segmented into distinct components based on voltage ranges. This segmentation simplifies the analysis by breaking down the complex current characteristics into manageable parts that can be analyzed independently. Each component can be processed with simpler algorithms, reducing overall computational complexity while maintaining high measurement precision.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent changes the parameter representation by dividing current data into different component values based on voltage thresholds. Instead of analyzing the raw current data directly, the method transforms it into component values that satisfy specific functions, making the defect detection process more systematic and less complex.

Inventive Principle:

Principle #35Parameter changes

3Measurement precision

If time data is grouped and redistributed using Weibull distribution, then defect categorization accuracy improves, but the computational processing time increases

Engineering Contradiction:

Improvedefect categorization accuracyVSAvoiddata processing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies parameter changes by fitting time data to Weibull distribution and extracting characteristic values. This statistical transformation converts raw time-to-breakdown data into standardized parameters that enable accurate defect categorization. By changing the data representation to distribution-based parameters, the method achieves high categorization accuracy while the computational overhead is managed through efficient statistical fitting algorithms.

Inventive Principle:

Principle #35Parameter changes

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

Accurately predicts defects, enhances wafer cutting and packaging accuracy, and improves semiconductor device reliability by categorizing defects and estimating product lifetime, ensuring reliability-certified products.

Implementation Method 1

dividing the current data into a first component value and a second component value using the current data and the data extract from within the decrease range; calculating a first quality index from the first component value satisfying a first function; and calculating a second quality index from the second component value satisfying a second function

Methodology Applied

Scientific EffectFowler-Nordheim tunneling:

Implementation Method 2

Methodology Applied

Scientific EffectTrap-assisted tunneling:

Implementation Method 3

dividing the time data into multiple groups; matching a probability distribution to a distribution of each of the plurality of groups; and redistributing the time data of the groups based on the current data

Methodology Applied

Scientific EffectWeibull distribution:

Data Source

PatentUS11668743B2Semiconductor device defect analysis method

Publication Date: 2023.06.06 SAMSUNG ELECTRONICS CO LTD

AI summary

A method of analyzing defects in a semiconductor device includes: collecting current data by applying a test voltage to the semiconductor device; extracting data within a decrease range from the current data; dividing the current data into a first component value and a second component value using the current data and the data extracted from within the decrease range; calculating a first quality index from the first component value satisfying a first function; and calculating a second quality index from the second component value satisfying a second function that is different from the first function.