Recent Advances in Microelectronics Reliability: Contributions from the European ECSEL JU project iRel40

Preface
	Acknowledgments
	Personal Acknowledgments
Contents
1 Reliability: Past and Present
	1.1 Physics of Degradation
	1.2 Robustness Validation
	1.3 The Fourth Wave
	References
2 Material Characterization and Modelling for FE-Based Reliability Assessment of PCBs and Electronic Systems
	Symbols and Abbreviations
	2.1 Introduction
	2.2 State of the Art in Material Models for PCB Simulation
	2.3 Characteristics of PCB Base Materials
		2.3.1 Overview of PCB Base Materials
		2.3.2 Dielectric Materials
			2.3.2.1 Epoxy Resin
			2.3.2.2 Glass Fibre-Reinforced Epoxy Resin
			2.3.2.3 Orthotropic Thermo-Mechanical Material Properties of Glass Fibre-Reinforced Epoxy Resin
			2.3.2.4 Orthotropic Coefficient of Thermal Expansion of Glass Fibre-Reinforced Epoxy Resin
			2.3.2.5 Further Specific Properties of Glass Fibre-Reinforced Epoxy Resin
			2.3.2.6 Material Model Validation and Quality Aspects of Dielectric PCB Material Models
		2.3.3 Copper
			2.3.3.1 Overview of Copper Types
			2.3.3.2 Thermo-Mechanical Properties of Copper
			2.3.3.3 Further Specific Properties of Copper
	2.4 Influence of Material Models on Simulation
		2.4.1 M2X Module
			2.4.1.1 Design and Stack-up of the M2X Module
			2.4.1.2 Numeric Simulation Model of the M2X Module
		2.4.2 Material Models for Simulation
			2.4.2.1 Material Models Based on Supplier Data Sheets and Literature
			2.4.2.2 Material Models Based on Temperature-Dependent, Orthotropic Measurements
		2.4.3 Material Model Validation
			2.4.3.1 Warpage Prediction by Numeric Simulation Compared to Measured Module Warpage
			2.4.3.2 Reliability Prediction by Numeric Simulation
	2.5 Suggestions for Comprehensive Material Models for Reliability Assessment of PCBs by FE Simulation
	2.6 Summary and Outlook
	References
3 Smart Optical Inline Metrology
	3.1 Introduction
	3.2 High-Dynamic Range Profilometry Using Spectral Imaging Interferometry
		3.2.1 DE-LCI Measurement Principle
			3.2.1.1 Experimental Approach
			3.2.1.2 Signal Model Formulation
			3.2.1.3 Fitting Procedure
		3.2.2 Resolution and Measurement Range
		3.2.3 Qualification and Measurement Results
			3.2.3.1 Height Evaluation and Repeatability
			3.2.3.2 HDR and 3D Measurements
			3.2.3.3 Tomographic Capabilities
		3.2.4 Inline Integration
		3.2.5 Conclusions
	3.3 Monitoring of Surface Material Decoating of Glass Carriers
		3.3.1 Introduction
		3.3.2 Experimental Setup
		3.3.3 Results
			3.3.3.1 Characterization of the Edge Region of the Glass Carrier
			3.3.3.2 Parameter Series for Optimal Decoating of the Metal Alloy Layer
			3.3.3.3 Process Monitoring and Spectrometric Analysis of the Decoating Process
		3.3.4 Conclusion
	3.4 Advanced Infrastructure for Statistical Data Analysis
		3.4.1 Introduction
		3.4.2 Declaration of the Application Frame
		3.4.3 From Measurement Setup to a Smart Sensor
		3.4.4 Infrastructure
			3.4.4.1 Scalable Architecture
			3.4.4.2 Infrastructure from the Logistical Point of View
			3.4.4.3 Data-Logistic Model in Continua Approximation
			3.4.4.4 Numerical Implementation of an Infrastructure Dynamic Model
			3.4.4.5 Model Calibration and Example Simulations
		3.4.5 Conclusion
	3.5 High-Power Broadband Light Sources in the Field of Optical Metrology
		3.5.1 Technical Broadband Light Sources with High Radiation Density
		3.5.2 State-of-the-Art High-Power Supercontinuum Light Sources
		3.5.3 Fiber Amplifier-Based Core-Pumped Supercontinua
		3.5.4 Supercontinuum Generation in the Fiber Cladding
		3.5.5 Numerical Simulation of Fiber Parameters
		3.5.6 Supercontinuum in the Core and Cladding
		3.5.7 Conclusion
	3.6 Conclusion
	References
4 Automated Classification of Semiconductor Defect Density SEM Images Using Deep Learning
	4.1 Introduction
	4.2 Data
		4.2.1 Data Source
		4.2.2 Data Collection
		4.2.3 Data Preparation
	4.3 Carinthia Dataset
		4.3.1 Dataset
		4.3.2 Model Development
		4.3.3 Results
	4.4 Madrid Dataset
		4.4.1 Dataset
		4.4.2 Model Development
		4.4.3 Deployment
	4.5 Conclusion and Outlook
	References
5 An Artificial Intelligence-Based Framework for Burn-in Reduction in the Semiconductor Manufacturing Industry
	5.1 Introduction
	5.2 Problem Statement and Formulation
	5.3 Methodology
		5.3.1 LSTM-Based CNN Method for Abnormal Conditions Detection
		5.3.2 PCA-OCSVM Method for Quality Prediction
			5.3.2.1 Feature Extraction Based on PCA
			5.3.2.2 Quality Prediction Based on OCSVM
		5.3.3 SVR-Based Method for Quality Prediction
			5.3.3.1 Feature Extraction Based on the Combination of PAA and PCA
			5.3.3.2 BI-Relevant Failure Prediction Based on SVR
	5.4 Applications
		5.4.1 Case Study 1: Validation of the Developed LSTM-Based CNN Method Using Production Process Data
		5.4.2 Case Study 2: Validation of the Developed PCA-OCSVM Method Using Wafer Map Data
		5.4.3 Case Study 3: Validation of the Developed SVR Method Using Electrical Test Data
	5.5 Concluding Remarks and Future Works
	References
6 Early Lifetime Estimation for Automotive LIDAR Using Realistic L4 Usage Profiles
	6.1 Introduction
	6.2 Testing Profiles
	6.3 Temperature Model for LIDAR
	6.4 Application: Burn-In Time Estimation
	6.5 Conclusion
	6.6 Future Work
	References
7 Improving the Reliability of Automotive Systems
	7.1 Introduction
		7.1.1 How Does the Automotive Industry Work? Nominations
		7.1.2 Electronification: A Disruption That Will Lead the Next Wave of Automotive Innovation
		7.1.3 Which Standards Must Automotive Suppliers Meet?
		7.1.4 The Increasing Complexity of In-vehicle Communications
		7.1.5 What to Expect from a Modern Product?
		7.1.6 In Summary
	7.2 A New Concept: From Mechanical Parts to Mechatronic Systems
		7.2.1 Digitalization: A Revolution in Mechanical Engineering
		7.2.2 Mechatronics: Two Worlds Converging in One System
		7.2.3 3D Printing and Simulation in Mechatronic Design
		7.2.4 In Summary
	7.3 Automotive Electronics: Challenges and Opportunities
		7.3.1 Standards and Guidelines for Safe and Reliable Components
			7.3.1.1 ISO 16750: Environmental Conditions and Electrical Loads
			7.3.1.2 ISO 26262: Functional Safety
			7.3.1.3 MISRA-C
			7.3.1.4 Automotive SPICE
		7.3.2 In Summary
	7.4 Cybersecurity: A Reliability Imperative
		7.4.1 Why Is Cybersecurity Important?
		7.4.2 Standards for Cybersecurity
		7.4.3 Cybersecurity: A Holistic Organizational Challenge
		7.4.4 In Summary
	7.5 Artificial Intelligence: Relying on a Black Box
		7.5.1 Training and Running AI Models
		7.5.2 Real-Time Computing at the Edge
		7.5.3 AI in Automotive Systems
		7.5.4 Standards for AI
		7.5.5 The Thin Line Between AI and Cybersecurity
		7.5.6 In Summary
	7.6 Discussion and Conclusions
		7.6.1 Challenges for the Industry
	References
8 Reliability Improvements for In-Wheel Motor
	8.1 Introduction
	8.2 In-Wheel Motor
	8.3 Description of the Measurement Devices
		8.3.1 Metrel Device
		8.3.2 ICM Device
		8.3.3 Testing Session
	8.4 Predictive Algorithm
		8.4.1 Prototype Design
		8.4.2 Tested Predictive Algorithms
	8.5 Experimental Environment
		8.5.1 Predictive Performance Results
			8.5.1.1 Cold Cycles
			8.5.1.2 Hot Cycles
			8.5.1.3 Combined Cold and Hot Cycles
	8.6 Conclusions and Future Work
	References
9 Big Data Streaming and Data Analytics Infrastructure for Efficient AI-Based Processing
	9.1 Introduction to Big Data
	9.2 Data Streaming
		9.2.1 Technologies and Tools for Data Streaming
			9.2.1.1 Data Ingestion Layer
			9.2.1.2 Message Queue and Stream Layer
			9.2.1.3 Stream Processing Layer
			9.2.1.4 Data Storage Layer
			9.2.1.5 Cluster Management Layer
			9.2.1.6 Output Layer
		9.2.2 Challenges and Considerations for Data Streaming
			9.2.2.1 Data Management and Processing
			9.2.2.2 Data Integrity, Security, and Governance
			9.2.2.3 Integration with Existing Systems
	9.3 Big Data and AI
		9.3.1 AI for Big Data Analytics and Processing
		9.3.2 Big Data and AI Applications
		9.3.3 The Impact of Big Data Quality on AI Models
	9.4 Monitoring and Visualization
		9.4.1 The Role of Visualization in Big Data and AI
		9.4.2 Tools for Big Data Visualization
		9.4.3 Task-Specific Visualization
			9.4.3.1 Over the Map Visualization
			9.4.3.2 Network Route Tracing and Performance Metrics
			9.4.3.3 Time-Series Analysis of Network Signals
	9.5 Case Studies
		9.5.1 Electric Motor Failure Detection and Diagnostic Approach Based on Motor Parameter Identification
		9.5.2  Big Data and AI Applications in Cybersecurity in Manufacturing
			9.5.2.1 The DECICE Project
	9.6 Future of Big Data and AI
	9.7 Conclusion
	References
10 An Outlook on Power Electronics Reliability and Reliability Monitoring
	10.1 Introduction
	10.2 Power Electronics Market and Failure Statistics from Field Experiences
	10.3 Power Electronics Degradation Mechanisms
		10.3.1 Chip-Related Degradation Mechanisms
			10.3.1.1 Electromigration
			10.3.1.2 Dielectric Contamination
		10.3.2 Package-Related Degradation Mechanisms
			10.3.2.1 Moisture Ingress
			10.3.2.2 Corrosion
			10.3.2.3 Thermomechanical Fatigue
	10.4 Power Electronics Reliability Monitoring
		10.4.1 Optical Thermal Inspection
			10.4.1.1 Thermoreflectance-Based Thermal Imaging
			10.4.1.2 Electroluminescence-Based Thermal Monitoring
		10.4.2 Thermal Test Chips
		10.4.3 Thermal-Pixel Test Chips
		10.4.4 Temperature-Sensitive Electrical Parameters
			10.4.4.1 On-State Resistance RDS(on)
			10.4.4.2 Source-Drain Voltage VSD Under Reverse Bias (Body Diode)
	10.5 Comparison of the Different Measurement Methods for Reliability Monitoring
	10.6 Conclusion
	References
11 Digital Twin Technology in Electronics
	11.1 Introduction
		11.1.1 What Is a Digital Twin?
		11.1.2 Simulation Models and How to Benefit from Reduced Order, Compact Simulation Models
		11.1.3 Basic Concept of Uncertainty Quantification
		11.1.4 Cross-Industry Interoperability
		11.1.5 Digital Twin as Used in Product Development/Management
	11.2 Digital Twin Use Cases
		11.2.1 Simulation-Based Digital Twin for a SO16 Current Measurement Device
		11.2.2 Modelling of a SiC MOSFET Inside a DC/DC Converter
	11.3 Conclusions
	References
12 A Framework for Applying Data-Driven AI/ML Models in Reliability
	12.1 Overview
	12.2 Reliability Challenges and AI/ML
	12.3 The iRel40 Framework
		12.3.1 Reliability Aspects
			12.3.1.1 Problem Type: Prognostics vs Diagnostics
			12.3.1.2 Reliability Assessment Methods
		12.3.2 Data and Machine Learning Aspects
			12.3.2.1 Model Output Data Characteristics
			12.3.2.2 Model Input Data Characteristics
			12.3.2.3 Machine Learning Approaches
	12.4 Discussion
	References
13 Health Monitoring Fatigue Properties of Solder Interconnects in LED Drivers
	13.1 Introduction
	13.2 Solder Characterization
		13.2.1 DMA Tests of Potting Compounds
		13.2.2 Effect of Temperature on Tensile and Creep Properties of Solder Materials
		13.2.3 DMA Curves of Potting Compounds
	13.3 Finite Element Simulations
	13.4 Prognostics and Health Monitoring
	13.5 Conclusions
	References
14 Executing Condition-Monitoring Algorithms on ARM Cortex-M4 Using TensorflowLite for Microcontrollers
	14.1 Introduction
	14.2 TensorFlow Ecosystem
	14.3 Experimental Setup
		14.3.1 Dataset
		14.3.2 Models
			14.3.2.1 CNN
			14.3.2.2 Transformer
			14.3.2.3 LSTM
		14.3.3 Hardware
	14.4 Experiments
		14.4.1 Testing TFLM Model Execution
		14.4.2 Memory and Execution Time Analysis
	14.5 Discussion and Conclusion
		14.5.1 Outlook
	References
15 Design Support for Reliable Integrated Circuits
	15.1 Introduction
	15.2 Technology and Device Selection Based on Reliability Data
	15.3 Verification of Transistor Degradation in Circuit-Level Design
	15.4 Case Study Applying RelXplorer and ReliaVision
	15.5 Summary
	References
16 Outlook to the Future of Reliability
	16.1 The (Reliability) Future Is Bright
	16.2 Applying Multi-scale and Multi-physics Simulations for Physics of Degradation
	16.3 Smarter Testing and Characterization
	16.4 ML/AI Embedding in Design for Reliability
	16.5 More Data, More PHM, and More Digital Twin
	16.6 Use Case: RUL Estimation for Electronic Devices
		16.6.1 Mission Profiles and Acceleration Factors
		16.6.2 Concept: Remaining Useful Life Estimation and System Status Assessment
		16.6.3 Use Case: Current Measurement Module
		16.6.4 Conclusions
	16.7 Final Remarks
	References
Index