Gallium nitride-based high-electron-mobility transistors (HEMTs) have demonstrated outstanding performance for high-power and high-frequency applications for defense and communication systems. However, there are many undesirable effects such as current collapse and increase in dynamic ON-resistance due to the surface donor states and the high polarization nature of the material [1], [2]. Nitride-based materials are more desirable for GaN passivation because the oxide-based materials have many oxygen contaminations on GaN. SiN has been proved as an effective passivation dielectric to reduce the surface states and can efficiently suppress current collapse in the GaN HEMTs [3]. However, the bandgap of SiN (∼5 eV) is not high enough to suppress leakage current. AlN has a large bandgap (∼6.2 eV) can effectively reduce leakage current as passivation layer [4]. In this work, we demonstrate GaN MIS-HEMT using AlN/SiN bilayer gate dielectric and passivation layer which combine the advantages of SiN and AlN. For comparing the performance of GaN MIS-HEMT with AlN/SiN bilayer gate dielectric, we also prepared reference devices with single layers dielectric. The wafer was divided into three samples after mesa and ohmic contact process. The dielectric and passivation layer were prepared differently for each sample: sample A with 5-nm SiN, sample B with 1-nm AlN/5-nm SiN, sample C with 5-nm AlN. Fig. 1 shows the I D-V D characteristics, where the drain current curves were measured at V G = 0 V. The SiN first passivated devices have higher I D and lower R ON as compared to AlN first passivated device. The R ON were 0.3 mΩ‧cm2, 0.3 mΩ‧cm2, and 0.33 mΩ‧cm2 for sample A, sample B, and sample C, respectively. For gate leakage shown in Fig. 2, GaN MIS-HEMTs using AlN as gate dielectric have lower gate leakage current. It means AlN can effectively suppress leakage current. Fig. 3 shows the percent degradation of R ON,DYN versus time. The OFF-state quiescent stress bias were set at V DSQ = 100 V and V GSQ = -10 V. The ON-state R ON,DYN is extracted at a gate voltage of 0 V and a drain voltage of 2 V. R ON,DYN degradation reflects the passivation quality including surface, interface, and barrier traps. GaN MIS-HEMT with AlN/SiN obviously suppresses R ON,DYN degradation compared to the single layer samples. Also, the SiN first passivated devices have slight R ON,DYN degradation. It indicates SiN has low interface trapping density on GaN surface and AlN can reduce trapping electrons charged through gate leakage. In conclusion, SiN was been proved that it has many good effects for GaN passivation such as decreasing of channel resistance, low surface state, low interface trapping density and low current collapse effect. AlN with high bandgap nature can suppress the leakage current. However, it would increase channel resistance, and cause severe current collapse effect. In this study, an effective AlN/SiN bilayer dielectric and passivation layer have been demonstrated for reducing current collapse effect and leakage current in GaN MIS-HEMT.Fig. 1 I D–V D characteristics of GaN HEMTs with different surface passivation layers.Fig. 2 Gate leakage current comparison in MIS-HEMTs structure.Fig. 3 Percent degradation of R ON,DYN versus time.[1] S. Binari, K. Ikossi, J. A. Rousos, et al., “Trapping effects and microwave power performance in AlGaN/GaN HEMTs,” IEEE Trans. Electron Devices, vol. 48, no. 3, pp. 465–471, Mar. 2001.[2] D. Jin and J. A. Alamo, “Methodology for the Study of Dynamic ON-Resistance in High-Voltage GaN Field-Effect Transistors,” IEEE Trans. Electron Devices, vol. 60, no. 10, pp. 3190–3196, Oct. 2013.[3] B. Green and K. Chu, “The effect of surface passivation on the microwave characteristics of undoped AlGaN/GaN HEMTs,” IEEE Device Lett., vol. 21, no. 6, pp. 268–270, Jun. 2000.[4] Z. Tang, S. Huang, Q. Jiang, S. Liu, C. Liu, and K. J. Chen, “High-Voltage (600-V) Low-Leakage Low-Current-Collapse AlGaN/GaN HEMTs With AlN/SiNx Passivation,” IEEE Device Lett., vol. 34, no. 3, pp. 366–368, Mar. 2013.
Search from 250 M+ research papersSearch from 250 M+ research papers
Apr 4, 2023
Chengcheng Zhang + 18
Open Access