Heteroepitaxial growth of In-face InN on GaN (0001) by plasma-assisted molecular-beam epitaxy

Abstract

The thermodynamic aspects of indium-face InN growth by radio frequency plasma-assisted molecular-beam epitaxy (rf-MBE) and the nucleation of InN on gallium-face GaN (0001) surface were investigated. The rates of InN decomposition and indium desorption from the surface were measured in situ using reflected high-energy electron diffraction and the rf-MBE “growth window” of In-face InN (0001) was identified. It is shown that sustainable growth can be achieved only when the arrival rate of active nitrogen species on the surface is higher than the arrival rate of indium atoms. The maximum substrate temperature permitting InN growth as a function of the active nitrogen flux was determined. The growth mode of InN on Ga-face GaN (0001) surface was investigated by reflected high-energy electron diffraction and atomic force microscopy. It was found to be of the Volmer–Weber-type for substrate temperatures less than 350°C and of the Stranski–Krastanov for substrate temperatures between 350 and 520°C. The number of monolayers of initial two-dimensional growth, in the case of Stranski–Krastanov mode, varies monotonically with substrate temperature, from 2 ML at 400°C to about 12 ML at 500°C. The evolution and coalescence of nucleated islands were also investigated as a function of substrate temperature. It was found that at higher temperature their coalescence is inhibited leading to porous-columnar InN thin films, which exhibit growth rates higher than the nominal value. Therefore, in order to achieve continuous InN layers on GaN (0001) a two-step growth approach is introduced. In that approach, InN is nucleated at low temperatures on GaN and the growth continues until full coalescence of the nucleated islands. Subsequently, this nucleation layer is overgrown at higher substrate temperature in order to achieve high-quality continuous films. The InN films grown by the two-step method were investigated by x-ray diffraction, Hall-effect measurements, and transmission electron microscopy. It was found that the lattice mismatch between InN and GaN is almost completely accommodated by the development of a misfit dislocation network at the interface. Optimum group-III to active nitrogen flux ratios and substrate temperature conditions were identified for the two-step growth process. Films, grown under those conditions, exhibited full width at half maximum of x-ray rocking curves at (0004) and (101¯5) diffractions equal to 360 and 435arcsec, respectively. Room-temperature Hall mobility was found to depend sensitively on the group-III to active nitrogen flux ratio during growth of the main step and to be independent of the structural properties of the films. Mobilities up to 860cm2∕Vs at carrier concentration of 1.6×1019cm−3 were measured.