PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
January 14, 2026AIP Advances0 citationsOpen Access

Enhanced surface passivation of AlGaN/GaN HEMTs via GaON-engineered interface for improved performance

View Full Paper
XWXiao WangUniversity of Science and Technology of ChinaYZYumin ZhangXidian UniversityLWLiang WangUniversity of Science and Technology of China

Key Points

  • This research aims to enhance the surface passivation of AlGaN/GaN HEMTs for improved performance and reliability.
  • Developed a novel multi-step surface passivation scheme.
  • Integrated various treatments including oxygen plasma and high-temperature annealing.
  • Utilized a dielectric layer formed by Al2O3 and SiO2.
  • Achieved a subthreshold swing of 66 mV/dec.
  • ION/IOFF ratio exceeded 109.
  • Demonstrated significant reductions in gate and off-state leakage.

Abstract

In this study, we report a novel multi-step surface passivation scheme for AlGaN/GaN high electron mobility transistors, integrating oxygen plasma treatment followed by high-temperature annealing (OPANT), Wet Buffered Oxide Etch (BOE) wet etching, remote plasma pre-treatment, and a stacked plasma-enhanced atomic layer deposition Al2O3/plasma-enhanced chemical vapor deposition SiO2 dielectric layer. This approach effectively forms a crystalline GaON layer at the interface, enhancing lattice compatibility with the dielectric and significantly reducing interface trap density. The proposed passivation yields a device with superior electrical performance, including a subthreshold swing of 66 mV/dec, an ION/IOFF ratio over 109, three orders of magnitude reduction in gate leakage and off-state leakage, and improved dynamic RON behavior under high drain bias stress. Devices incorporating a GaON gate interlayer exhibit enhanced gate reliability.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Wang et al. (2026) studied this question.

synapsesocial.com/papers/6966e71813bf7a6f02bff65ehttps://doi.org/10.1063/5.0313966
Ask AI
Helpful
Bookmark
Share
View Full Paper