Optical study of gallium and nitrogen polarity layers of GaN grown on sapphire

Abstract

A confocal Raman spectroscopic study was carried out on either side of an intentionally grown GaN inversion domain boundary between a pair of strips with opposite (Ga-or N-) polarity. It is shown that the Raman spectra on the N-polarity side displays an A 1 (TO) mode, prohibited by symmetry considerations, meanwhile on Ga-polarity material this peak is absent, indicating a lower density of defects present in this region. The Raman spectra reveal that in the lateral direction, the change in structural quality accross the inversion domain boundary is rather continuous and extends along 4 AE 1 mm.

Fabrication of lateral lattice-polarity-inverted GaN heterostructure

“…3(a) shows the polarized m-Raman spectra for the two representative domains: GaN on the Ga-polarity template (domains A) and that on the FIB-processed region (domain B (iv), with a nominal process depth of 68 nm in this case). The presence of the selection rule of A 1 ðLOÞ mode and the absence of either its noticeable damping or other forbidden modes show their high structural perfections in both domains, contrary to the previous reports [3,4]. As shown in Fig.…”

“…Also the surface potential difference of 1.17 eV is the largest value ever reported, showing that the density of the microscopic inversion domain within the individual domains found to be small, and that the background sheet carrier density here is $3 Â 10 12 cm À2 , which is not enough to compensate the polarization induced charges of $2:1 Â 10 13 cm À2 , as totally different from the previous reports where the polarization charges have been completely screened by the background carriers [3,4]. And the reason for the intermediate value of the surface potential difference, like 0.24 eV for a domain B (ii) with a process depth of 43 nm, is interpreted as a result of the mixed polarity due to the nonflatness of the FIB processing surface: sapphire surface was partly exposed but Ga-polarity template was also left partially before the regrowth.…”

“…The estimated carrier concentrations for representative domains A and B (iv) are around 3.3-4:8 Â 10 16 and $5:2 Â 10 16 cm À3 , respectively. This fact indicates the low carrier concentrations in either domain of our sample, as opposed to the report that the N-polarity domains always suffered from a severe degeneracy of the background electrons [3,4]. Furthermore the maximum allowable offset of the Fermi level for domain B (iv) with respect to A is only þ12 meV.…”

“…But until recently, as Park and Cros pointed out [3,4], previously reported polarity-inverted GaN suffered from significant degradation of the crystal quality in the N-polarity part of domains, thus it must be admitted that its fabrication process is not fully established yet. Therefore in this work, we demonstrate the improvement in the crystal quality of the polarity-inverted GaN heterostructure, which was fabricated by a patterning of a thin Ga-polarity template on a sapphire (0 0 0 1) substrate and successive regrowth.…”