Wet thermal oxidation of GaN

Journal of Electronic Materials, Mar 1999 by Readinger, E D, Wolter, S D, Waltemyer, D L, Delucca, J M, Et al

Thermal oxidation of GaN was conducted at 700-900degC with O2, N2, and Ar as carrier gases for 525-630 Torr of H20 vapor. Upon oxidation of both GaN powders and n-GaN epilayers, the monoclinic beta-Ga^sub 2^O^sub 3^ phase was identified using glancing angle x-ray diffraction. The chemical composition of the oxide was verified using x-ray photoelectron spectroscopy. In experiments conducted using GaN powder, the oxide grew most rapidly when O2 was the carrier gas for H20. The same result was obtained on n-type GaN epilayers. Furthermore, the thickness of the oxide grown in H20 with O2 as the carrier gas was found to be proportional to the oxidation time at all temperatures studied, and an activation energy of 210+/-10 kJ/mol was obtained. Scanning electron microscopy revealed a smoother surface after wet oxidation than was reported previously for dry oxidation. However, cross-sectional transmission electron microscopy revealed that the wet oxide/GaN interface was irregular and non-ideal for device fabrication, even more so than the dry oxide/GaN interface. This observation was consistent with poor electrical properties.

Key words: GaN, thermal oxidation

INTRODUCTION

Group III nitrides and their alloys are under rapid development for optoelectronic devices in the blue, green, and ultraviolet (LTV) portion of the spectrum, with applications including data storage, displays, and detectors. They are also receiving attention for high power and high temperature electronics. While much work has been done to study the native thermal oxides on semiconductors such as Si and more recently SiC and the Al-bearing III-V semiconductors, very few studies have been conducted on the thermal oxidation of GaN. Wolter et al. studied the oxidation of GaN in dry air1 which resulted in the formation of a polycrystalline oxide with a rough surface.2 It was speculated that wet oxidation of GaN might result in the formation of an oxide with improved polycrystalline morphology or even an amorphous oxide. Therefore, in this study, we have investigated the thermal oxidation of GaN in H20 vapor and examined the effect of different carrier gases upon the oxidation process. The electrical properties of the thermal oxides are also evaluated.

PROCEDURE

GaN powder (-325 mesh) with a purity of at least 99.99% (metals basis) was oxidized in H20 vapor for 5 h at 700-900degC using ultra-high purity (UHP) O2 N2, and Ar as carrier gases. Oxidation of samples in dry O2 at the same temperatures was also performed for comparison. The powder samples were characterized using a Phillips multipurpose diffractometer, and 0-20 scans were performed using CuK^sub a^ radiation. The relative peak intensities were compared for similar oxidation conditions to give qualitative information about oxide growth rates.

GaN epitaxial layers were grown on (0001) sapphire using a vertical hydride vapor phase epitaxy process3 with a sputtered ZnO pretreatment. The n-type films were grown at 1050degC for one hour at a rate of 15 gm/h and had a carrier concentration of 2.2 x 10^sup 17^ cm^sup -3^. The samples were rinsed in acetone, methanol, and deionized (DI) water, after which they were placed in a 1:1 solution of HCl:DI water for 60 s and rinsed in DI water.

The samples were placed in a quartz boat within a quartz tube, and then the tube was purged with the carrier gas for at least 20 min before it was placed in a horizontal furnace. Once the tube was placed into the furnace, gases were diverted into the previously by-passed bubbler. Using a thermocouple inserted in the quartz tube, we determined that the time for the sample to reach the intended temperature did not exceed 6 min. The carrier gases were maintained at a flow rate of 25 sccm through a bubbler containing DI water heated to 93+/-1degC. Oxidation was performed at 700-900degC for 1-50 h, and results were compared to a previous study of the oxidation of GaN in dry air.1,2,4 Following oxidation, the quartz tube was removed from the furnace and allowed to cool in air to room temperature before the sample was removed.

A consistent volume of DI water was used in the bubbler before each experiment, which has been reported to increase the reproducibility of the partial pressure of water vapor transferred to the furnace.5 In addition, heating tape maintained at approximately 110C was used on the gas feed line from the bubbler to the tube furnace to minimize condensation between the bubbler and furnace. The atomic percent of H20 vapor present in the carrier gas was estimated to be 77+/-8% for any given furnace temperature and carrier gas combination. These calculations were based on the flow rate of the gas and the volume of H20 lost from the bubbler.

The initial identification of the thermal oxide was performed using glancing angle x-ray diffraction (XRD) using CuK^sub a^ radiation. A Kratos analytical XSAM 800 pci system with a MgK^sub a^ x-ray source was used for xray photoelectron spectroscopy (XPS) with a chamber base pressure of 1 x 10^sup -9^ Torr. The Ga 3d, N ls, and O ls core level peaks were chosen for study to further corroborate the oxide identification.16Further analysis was performed using a Phillips XL-20 scanning electron microscope (SEM) to examine the oxide surface using a secondary electron detector, and the oxidized GaN epilayers were examined in cross section after cleaving the sapphire wafer. These cross-sectional samples were used to measure layer thicknesses in order to extract data on growth kinetics. The thermal oxide was also imaged at greater magnifications using cross-sectional transmission electron microscopy (TEM) on samples prepared by a focused ion beam technique.7 Both wet and dry oxidation conditions were compared using a Philips 420T TEM.