In this article, the growth and coalescence of semi‐polar false(11true2‾2false) oriented GaN layers, deposited on pre‐structured r‐plane sapphire substrates, is studied with the help of Si‐doped marker layers. It has been found to be very important to adjust the shape of the initial GaN stripes by varying the growth temperature to obtain not only a smooth surface, but also a small density of basal plane stacking faults (BSFs) and threading dislocations (TDs) on the wafer surface. With the help of transmission electron microscopy (TEM) and cathodoluminescence measurements (CL), we can conclude that during growth, we need to achieve a compromise between small BSF density, small TD density, and perfect coalescence with smooth surface, free of fissures, and other growth artifacts. Also the formation of arrow‐head‐shaped surface artifacts called “chevrons” can be understood to be caused by imperfect coalescence. We observe with the help of the marker layers that the growth rate fluctuates between neighboring stripes. This effect strongly increases for higher growth temperature.

We demonstrate the application of low-temperature cathodoluminescence (CL) with high lateral, depth, and spectral resolution to determine both the lateral (i.e., perpendicular to the incident primary electron beam) and axial (i.e., parallel to the electron beam) diffusion length of excitons in semiconductor materials. The lateral diffusion length in GaN is investigated by the decrease of the GaN-related luminescence signal when approaching an interface to Ga(In)N based quantum well stripes. The axial diffusion length in GaN is evaluated from a comparison of the results of depth-resolved CL spectroscopy (DRCLS) measurements with predictions from Monte Carlo simulations on the size and shape of the excitation volume. The lateral diffusion length was found to be (95 ± 40) nm for nominally undoped GaN, and the axial exciton diffusion length was determined to be (150 ± 25) nm. The application of the DRCLS method is also presented on a semipolar (112¯2) sample, resulting in a value of (70 ± 10) nm in p-type GaN.

No abstract

Spatially resolved luminescence and Raman spectroscopy investigations are applied to a series of (112¯2)-GaN samples grown by hydride vapor phase epitaxy (HVPE) grown over an initial layer deposited by metal organic vapor phase epitaxy on patterned sapphire substrates. Whereas these two differently grown GaN layers are crystallographically homogeneous, they differ largely in their doping level due to high unintentional oxygen uptake in the HVPE layer. This high doping shows up in luminescence spectra, which can be explained by a free-electron recombination band for which an analytical model considering the Burstein-Moss shift, conduction band tailing, and the bandgap renormalization is included. Secondary ion mass spectrometry, Raman spectroscopy, and Hall measurements concordantly determine the electron density to be above 1019 cm−3. In addition, the strain state is assessed by Raman spectroscopy and compared to a finite element analysis.

No abstract