Comprehensive analysis of gate stack barrier resistance in GaN-based MIS-Hemts Using Lumped Element Modelling

Abstract

Gallium nitride (GaN)-based metal-insulator-semiconductor high-electron-mobility transistors (MIS-HEMTs) have emerged as a pivotal technology in high-power and high-frequency applications due to their superior electrical and thermal properties. However, optimizing gate stack barrier resistance remains a critical challenge, influencing key device parameters such as gate leakage current, barrier resistance, and thermal stability. This study aims to address these challenges by developing a comprehensive lumped element model to analyze gate stack barrier resistance and propose strategies for optimizing device performance. Experimental analyses were conducted on GaN MIS-HEMTs with varying Al₂O₃ dielectric thicknesses (10–30 nm), and their electrical and thermal characteristics were measured using advanced semiconductor device analyzers. The lumped element model was employed to simulate and predict the behavior of gate stack parameters, validated by statistical tools such as ANOVA and regression analysis. The results revealed that increasing dielectric thickness significantly reduces gate leakage current density, with devices featuring a 30 nm dielectric exhibiting the lowest leakage of 2.8×10-8 A/cm2. Barrier resistance showed a linear increase with thickness (R2=0.98), while gate capacitance decreased proportionally. Devices maintained thermal stability up to 150 ∘C, with minimal performance degradation. These findings highlight the importance of optimizing dielectric thickness, with an optimal range of 20–25 nm identified to balance performance metrics. The study concludes that incorporating advanced high-k materials, precise deposition techniques, and interface engineering can further enhance device reliability and efficiency. By integrating experimental insights with predictive Modelling, this research provides actionable recommendations for the development of next-generation GaN-based MIS-HEMTs.