研究目的
To analyze two important reliability issues in AlGaN/GaN devices: positive bias temperature instability (PBTI) and time-dependent dielectric breakdown (TDDB), focusing on the effects of different dielectric materials and device architectures.
研究成果
The study concludes that Gate-SiO2 devices show charge trapping in both dielectric and interface, with a universal trapping rate behavior. High-κ dielectrics exhibit faster trapping in existing traps with negligible interface state generation. Adding an AlN layer increases trap density and slows charge release. For TDDB, double GET structures with thick passivation show narrower Weibull distribution and longer time to failure, making them more suitable for high-power and high-temperature applications.
研究不足
The methodology for PBTI is only useful for moderate degradation and when characterization does not induce additional stress. The study may be limited by device variability and the specific materials and structures used. Extrinsic factors like buffer limitations in earlier TDDB studies were addressed but could still affect results.
1:Experimental Design and Method Selection:
The study involves PBTI analysis in MOS-HEMTs and TDDB analysis in GET-SBDs. For PBTI, DC characterization and fast pulse techniques are used depending on the dielectric. For TDDB, constant voltage stress (CVS) is applied with monitoring of current variation.
2:Sample Selection and Data Sources:
Devices are fabricated by STMicroelectronics and imec. MOS-HEMTs have different gate dielectrics (SiO2, Al2O3, AlN/Al2O3). GET-SBDs have single or double GET structures with thin or thick passivation layers.
3:3). GET-SBDs have single or double GET structures with thin or thick passivation layers.
List of Experimental Equipment and Materials:
3. List of Experimental Equipment and Materials: Equipment includes devices with specific dielectrics and structures, as detailed in Table I. Materials include SiO2, Al2O3, AlN, Si3N4, and metallizations like Ni and Ti/Al.
4:Experimental Procedures and Operational Workflow:
For PBTI, initial stabilization with negative bias, followed by stress and recovery phases with voltage and temperature variations. For TDDB, reverse bias sweeps and CVS with high voltage and temperature to accelerate degradation.
5:Data Analysis Methods:
Data is analyzed using models like universal decreasing behavior for trapping rate, saturating log-time dependence for high-κ dielectrics, and Weibull distribution for TDDB. Parameters such as threshold voltage shift and time to breakdown are extracted.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容