Search

In this paper, we carried out an analysis of the breakdown characterization method by the investigation on off-state leakage currents and breakdown curves. For conventional breakdown, seven kinds of breakdown curves are summarized and it is found that only two of them can be shown by the conventional three-terminal breakdown characterization method reasonably. For the other five kinds of breakdown curves, the value of the gate leakage current is larger than that of the drain leakage current for a certain drain-bias range. Besides, the source leakage current cannot represent the buffer leakage current, and the values and signs of them are different. These problems contradict the conventional characterization method, indicating that the conventional method should be modified to characterize the breakdown mechanisms correctly. The similar problems also exist for time-dependent breakdown. The buffer and drain-gate leakage currents were obtained by a simple method and the conventional breakdown characterization method was modified by using these two currents. The problems in the conventional breakdown characterization method are solved by using the modified method. Experiments indicate that the modified breakdown characterization method is crucial to investigate the breakdown mechanisms, especially when source-gate leakage current cannot be neglected compared with the buffer leakage current. [ABSTRACT FROM AUTHOR]

Frequency-dependent conductance measurements were carried out to investigate the trap states induced by reactive ion etching in AlGaN/GaN high-electron-mobility transistors (HEMTs) quantitatively. For the non-recessed HEMT, the trap state density decreases from 2.48 × 1013 cm−2·eV−1 at an energy of 0.29 eV to 2.79 × 1012 cm−2·eV−1 at ET = 0.33 eV. In contrast, the trap state density of 2.38 × 1013–1.10 × 1014 cm−2·eV−1 is located at ET in a range of 0.30–0.33 eV for the recessed HEMT. Thus, lots of trap states with shallow energy levels are induced by the gate recess etching. The induced shallow trap states can be changed into deep trap states by 350 °C annealing process. As a result, there are two different types of trap sates, fast and slow, in the annealed HEMT. The parameters of the annealed HEMT are ET = 0.29–0.31 eV and DT = 8.16 × 1012–5.58 × 1013 cm−2·eV−1 for the fast trap states, and ET = 0.37–0.45 eV and DT = 1.84 × 1013 − 8.50 × 1013 cm−2·eV−1 for the slow trap states. The gate leakage currents are changed by the etching and following annealing process, and this change can be explained by the analysis of the trap states. [ABSTRACT FROM AUTHOR]

In this paper, we present the combination of drain field plate (FP) and Schottky drain to improve the reverse blocking capability, and investigate the reverse blocking enhancement of drain FP in Schottky-drain AlGaN/GaN high-electron mobility transistors (HEMTs). Drain FP and gate FP were employed in a two-dimensional simulation to improve the reverse blocking voltage (VRB) and the forward blocking voltage (VFB). The drain-FP length, the gate-FP length and the passivation layer thickness were optimized. VRB and VFB were improved from −67 V and 134 V to −653 V and 868 V respectively after optimization. Simulation results suggest that the combination of drain FP and Schottky drain can enhance the reverse blocking capability significantly. [ABSTRACT FROM AUTHOR]