Plasma-assisted molecular beam epitaxy of GaN on porous SiC substrates with varying porosity

Abstract

Abstract:We have grown GaN on porous SiC substrates and studied the effect of substrate porosity on the overgrown film quality in terms of defect structure and density and film strain. The growth was performed by plasma-assisted molecular beam epitaxy (PAMBE). The GaN films were characterized by x-ray, transmission electron microscopy (TEM) and wafer curvature measurements by surface profilometry. TEM images show that the GaN film grown on porous substrates contains open tubes and a low dislocation density in regions between tubes. We discuss various growth mechanisms that can lead to these defect features in the GaN film. However, we do not find any overall improvement in the x-ray rocking curve FWHM of the GaN films grown on porous substrates compared to those on nonporous substrates. It was found that the GaN films grown on porous SiC were significantly more strain relaxed compared to those grown on nonporous substrate. We propose various mechanisms that can lead to the reduction in strain in GaN films grown on porous substrates and compare the data with finite element analysis (FEA) simulations of such a system.

Combined MOCVD and MBE growth of GaN on porous SiC

“…Additionally, some lateral overgrowth of GaN was found to occur over SiC pores larger than ∼200 nm, and in regions between tubes the density of threading dislocations was found to be relatively low. Nevertheless, there was no overall improvement in the dislocation density of the GaN films on porous compared to nonporous substrates [7].…”

“…1(b) was prepared by removing the top ~200 nm from the porous sample by reactive ion etching (RIE) in SF 6 gas for 4 minutes and then H-etching for 50 sec at 1650°C. This combination of H-etching and RIE was used to produce a variation in surface pore density, as described elsewhere [7,8], with resulting densities of 4.6 µm for the surfaces shown in Figs. 1(a) and (b), respectively.…”

“…Earlier studies have shown improved GaN film quality on porous SiC [2][3][4]. It was recently shown that the GaN grown by molecular beam epitaxy (MBE) on porous SiC is considerably more strain relaxed than the GaN grown on nonporous SiC [7]. It was found for porous substrates that the MBE GaN films contain open tubes originating from the pores in the SiC substrate and extending up to the surface of the film [7].…”

“…It was recently shown that the GaN grown by molecular beam epitaxy (MBE) on porous SiC is considerably more strain relaxed than the GaN grown on nonporous SiC [7]. It was found for porous substrates that the MBE GaN films contain open tubes originating from the pores in the SiC substrate and extending up to the surface of the film [7]. These tubes formed in the MBE-grown GaN because of insufficient lateral growth over the substrate pores.…”

“…In the present study, we have grown GaN by metal organic chemical vapor deposition (MOCVD) on top of MBE-grown GaN films prepared as in our prior work [7]. The aim of the experiment was to use the lateral growth in MOCVD to close-up the open tubes observed in the MBE GaN on porous SiC and, in the process of doing so, perhaps to reduce the dislocation density in the overgrown film.…”