Description
Date/Time: 04-23-2019 - Tuesday - 05:00 PM - 07:00 PM
Shuo Wang1 Katia March1 Peter Rez1 Fernando Ponce1

1, Arizona State University, Tempe, Arizona, United States

AAlthough there are many techniques that can detect bandgap states associated with point defects in the lattice, it is not routinely possible to determine the type of defect at submicron spatial resolution. Here we show that high-resolution electron energy loss spectroscopy (EELS) in a scanning transmission electron microscope can locate and identify point defects with a sub-nanometer probe in a wide-bandgap BAlN semiconductor.
Two BAlN thin films of 50 nm thickness were grown by metalorganic chemical vapor deposition on top of AlN templates, with B/(B+Al) gas-flow ratios of 0.12 and 0.18, respectively. In the atomic-resolution high-angle annular dark-field images, the two films exhibit considerable variation in the B concentration. The high-resolution monochromated spectra from the high B concentration BAlN film taken at 60kV show energy losses in the bandgap region. The film with low B concentration has similar energy thresholds, but with lower intensities. These energy-loss thresholds are still detectable in the AlN substrate, although with diminishing intensity as the beam is moved further from the interface, due to the long-range nature of the electromagnetic interaction. Since there is an analytical theory for the spatial variation of the signal as a function of distance between the probe and the dipole, it would be in principle possible to fit this functional form to the observed signal to localize the dipole with nanometer resolution.
To identify the origin of these features, calculations of densities of states with the VASP Density Functional Theory (DFT) code were used to explore whether postulated point defects give rise to states in the band gap, and whether transitions involving these states match experimental observations. VASP DFT calculations were performed using projection augmented wave, local-density approximation potentials for single Al, B and N interstitial atoms, single Al and N vacancies in a supercell constructed from 3x3x2 AlN unit cells. Although the bandgap energy is underestimated as 4.6 eV, it could be argued that the DFT energies for the point defect states are reliable.
In our calculations, there are no bandgap states when B is substituted for Al, and the density of states is very similar to that of AlN. In all the other cases the calculations showed states in the band gap, consistent with previous reports on point defects in AlN. The EELS thresholds at 0.39 eV and 0.79 eV respectively, are attributed to B interstitials. It is expected that they would be present in both the high and low concentration B films, albeit with lower intensity in the lower concentration film, given that the solubility limit of B in AlN is reportedly 2.8%. It would seem that other thresholds at 0.53 eV and 0.6 eV, arise from N or Al vacancies, and possibly Al interstitials. Not surprisingly these thresholds attributed to displaced Al and N atoms are more likely in the film with higher B concentration.

Meeting Program