C. S. Oh

Conductive layer near the GaN/sapphire interface and its effect on electron transport in unintentionally doped n-type GaN epilayers

Temperature-dependent Hall effect measurements on unintentionally doped n-type GaN epilayers show that, above room temperature, the Hall-mobility values of different samples vary parallel with each other with temperature. We demonstrate that this anomaly is mainly due to a conductive layer near the GaN/sapphire interface for thin samples with low carrier density. Through trapping electrons, threading edge dislocations (TEDs) debilitate the epilayer contribution in a two-layer mixed conduction model involving the epilayer and the near-interface layer. The trapping may, in part, explain low mobility and anomalous transport in pure GaN layers. Scattering by TEDs is important only at low temperatures.

Effects of growth rate of a GaN buffer layer on the properties of GaN on a sapphire substrate

We studied the effects of the growth rate of a GaN buffer layer grown on a GaN epilayer. It was found that this growth rate plays a key role in improving the quality of the GaN film on a sapphire substrate and an optimum growth rate exists that yields the best crystal quality. A GaN film grown on a buffer layer with the optimum growth rate of 18.3 nm/min has an electron Hall mobility of 539 cm2/V s and a dislocation density of approximately 2×108 cm−2. These improvements of GaN film qualities are illustrated by the promotion of the lateral growth mode.

Bias-assisted photoelectrochemical oxidation of n-GaN in H2O

Growth of gallium oxide on n-GaN was realized in H2O by bias-assisted photoelectrochemical (PEC) oxidation using Al as a counterelectrode instead of a Pt commonly used in the PEC process. Although the growth of the oxide was not observed at below 2 V, the initial oxide growth rate of 8.7 nm/min was shown at a bias of 15 V and ultraviolet light intensity of 300 mW/cm2. However, the growth rate lowered and oxide thickness was saturated to 340 nm. The saturated oxide thickness and initial growth rate were increased with the applied bias. The homogeneous oxide growth and near stoichiometric composition of Ga2O3 were observed in Auger electron spectroscopy analysis results.

AlGaN∕GaN metal-oxide-semiconductor heterostructure field-effect transistor with oxidized Ni as a gate insulator

We fabricated the AlGaN∕GaN metal-oxide-semiconductor heterostructure field-effect transistor (MOSHFET) using the oxidized Ni(NiO) as a gate oxide and compared electrical properties of this device with those of a conventional AlGaN∕GaN heterostructure field-effect transistor (HFET). NiO was prepared by oxidation of Ni metal of 100A at 600°C for 5min in air ambient. For HFET and MOSHFET with a gate length of 1.2μm, the maximum drain currents were about 800mA∕mm and the maximum transconductances were 136 and 105mS∕mm, respectively. As the oxidation temperature of Ni increased from 300 to 600°C the gate leakage current decreased dramatically due to the formation of insulating NiO. The gate leakage current for the MOSHFET with the oxidized NiO at 600°C was about four orders of magnitude smaller than that of the HFET. Based on the dc characteristics, NiO as a gate oxide is comparable with other gate oxides.

Thermal distributions of surface states causing the current collapse in unpassivated AlGaN∕GaN heterostructure field-effect transistors

The dc characteristics of the AlGaN∕GaN heterostructure field-effect transistors were examined at temperatures ranging from 25 to 260 °C under white light illumination. Drain current collapse measured was defined by the difference of drain current between light on and light off at Vgs=1V and Vds=5V. The surface-passivated device showed no drain current collapse, but the unpassivated device showed severe drain current collapse at 25 °C. Drain current and drain current collapse with an increase in temperature reduced, which resulted from the reduction of the electron mobility or saturation velocity and the thermal activation of the trapped electrons, respectively. Eventually, drain current collapse disappeared completely above 250 °C. The behavior of the temperature-dependent drain current collapse showed that the surface states for trapping electrons were continuously distributed with the temperature not having specific energy states.

Co-doping characteristics of Si and Zn with Mg in p-type GaN

The authors investigated the doping characteristics of Mg doped, Mg-Si co-doped, and Mg-Zn co-doped Gan films grown by metalorganic chemical vapor deposition. They have grown p-GaN film with a resistivity of 1.26 {center_dot} cm and a hole density of 4.3 x 10{sup 17} cm{sup {minus}3} by means of Mg-Si co-doping technique. The Mg-Si co-doping characteristic was also explained effectively by taking advantage of the concept of competitive adsorption between Mg and Si during the growth. For Mg-Zn co-doping, p-GaN showing a low electrical resistivity (0.7 {center_dot} cm) and a high hole concentration (8.5 x 10{sup 17} cm{sup {minus}3}) was successfully grown without the degradation of structural quality of the film. Besides, the measured specific contact resistance for Mg-Zn co-doped GaN film is 5.0 x 10{sup {minus}4} {center_dot} cm{sup 2}, which is lower value by one order of magnitude than that for only Mg doped GaN film (1.9 x 10{sup {minus}3} {center_dot} cm{sup 2}).

Allowable Substrate Bias for the Etching of n‐GaN in Photo‐Enhanced Electrochemical Etching

Etching characteristics for n-GaN on bias voltage were examined in various chemical solutions including the solutions not considered as etchants, n-GaN was not etched in HNO 3 and CH 3 COOH solutions during photo-enhanced electrochemical (PEC) etching under UV illumination of 90 mW/cm 2 . It was successfully etched in 1% CH 3 COOH solution by applying a bias larger than 0.6 V and the etch rate was increased to 23 nm/min at 1.5 V. Also, etching of n-GaN began to appear at 0.4 V in 0.05% HNO 3 solution. On the contrary, n-GaN was etched at a rate of 12 nm/ min in 0.04M KOH solution and 6 nm/min in 0.01M H 3 PO 4 solution even though the bias was not applied and -0.4 and -0.3 V were needed to stop the etching of n-GaN, respectively. The increase of etch rate on the reverse bias was shown for all the examined solutions and the critical bias for n-GaN etching was not varied on the UV illumination intensity.