Ab initio calculations of phonon dispersion relations for bulk and surface of cubic InN

Overcoming the Miscibility Gap of GaN/InN in MBE Growth of Cubic InxGa1–xN

“…The origin of additional broad Raman signals, such as the second signal observed here in close vicinity to the LO-like mode of the alloy, is often second-order Raman scattering . However, in our case, it seems unlikely that second-order Raman scattering is the origin of the additional peaks seen here, based on the phonon dispersions calculated for c-GaN and c-InN. , In our series of samples, preliminary experimental evidence suggests that both effects (i.e., the strong broadening of the LO-like phonon signal at high x (In) and the occurrence of a second signal at lower x (In)) may be related to short-range CuPt-type ordering, which yields alternating In- and Ga-rich lattice planes along the [111] direction. This phenomenon is well-known from Ga 1– x In x P alloys with intermediate x (In). , The ideal structure that corresponds to the CuPt-type ordering is a rhombohedral crystal structure, where the cubic [111] direction acts as a high symmetry three-fold rotational axis and there is a basis of four atoms (2 N, 1 Ga, 1 In) per unit cell instead of two atoms, as in zincblende crystals.…”

“…64 However, in our case, it seems unlikely that second-order Raman scattering is the origin of the additional peaks seen here, based on the phonon dispersions calculated for c-GaN and c-InN. 65,66 In our series of samples, preliminary experimental evidence suggests that both effects (i.e., the strong broadening of the LO-like phonon signal at high x(In) and the occurrence of a second signal at lower x(In)) may be related to short-range CuPt-type ordering, which yields alternating In-and Ga-rich lattice planes along the [111] direction. This phenomenon is wellknown from Ga 1−x In x P alloys with intermediate x(In).…”

“…This means that two LO-like phonons (A 1 symmetry in case of the corresponding rhombohedral structure) become allowed in the presence of CuPt-type ordering. Furthermore, the theoretical work that has been done on phonon dispersions of c-GaN and c-InN 65,66 reports that the LO phonon dispersion of c-InN between Γ and L is flatter than that of c-GaN. This suggests that two LOlike Raman signals are observable at lower In content (as indicated by the arrows in Figure 6) but then merge into one broader feature at higher In content.…”

“…The frequency difference calculated based on the values for c-GaN is in reasonable agreement with the frequency difference between the two Raman features observed in the spectra. Although the transverse (TO) modes of zincblende materials are forbidden in the backscattering from (100) surfaces, corresponding modes can be discerned in the spectra of the binary materials at 551 cm −1 (c-GaN 60,61,69 ) and 458 cm −1 (c-InN 62,65,66 ). No clear signals can be discerned for the alloy thin films; this is because the anticipated Raman signals are covered by the strong signals of the LO phonon of the Si substrate and the TO mode of the c-GaN buffer layer.…”