Localization of excitation in InGaN epilayers

Abstract

Energy scalability of the excitation–emission spectra of InGaN epilayers, quantum wells and light-emitting diodes provided indirect evidence for a fundamental common cause of the remarkable optical properties of this commercially important semiconductor alloy. Phase segregation on the nanoscale (accidental quantum dot formation) has generally been accepted as the mechanism of the spectral energy scaling (K.P. O’Donnell, R.W. Martin and P.G. Middleton, Phys. Rev. Lett. 82 237 (1999)). Recently, however, the downsizing of the InN bandgap, from 2 to about 1 eV, has prompted a re-examination of the observations. Here, we present new structural evidence of InGaN nanostructure, obtained from a comparative analysis of Ga and In K-edge EXAFS (extended X-ray absorption fine structure) of a wide range of In x Ga1− x N epilayer samples. The mean In–Ga and Ga–In next-nearest-neighbour (NNN) separations are found to be unequal in length for InN-poor (0.1 < x < 0.4) samples. The degree of inequality increases with decreasing InN fraction, x, and therefore correlates with luminescence efficiency in this range of alloy composition. We propose that the breakdown of In/Ga randomicity in InGaN alloys is associated with efficient excitation–emission in blue-green light-emitting devices. Although non-randomicity may lead to a weak quasi-localization of excitation, through the suppression of energy back-transfer, the issue of strong exciton localization in InGaN is not directly addressed by these results.