Assessment of the short and long range homogeneity of MOVPE-grown InGaN epilayers R. Fornari 1) , M. Bosi 1) , M. Avella 2) , O. Martinez 2) , J. Jimenez 2) 1) Istituto IMEM - CNR, Area delle Scienze 37/A, 43010 Fontanini, Parma, Italy , 2) Dip. Fisica de la Materia Condensada, Universidad de Valladolid, 47011 Valladolid, Spain Abstract InGaN epilayers grown by MOVPE were investigated by micro-Raman spectroscopy, photoluminescence and spectrally-resolved cathodo-luminescence in order to analyse the indium distribution at the micro and macro scale. It was found that the In molar fraction increases from centre to edge of the 2" wafers. The Raman spectra of the ternary alloys exhibit additional In-related modes, with respect to GaN spectra. One of these Raman bands is tentatively ascribed to In clustering, further justified by CL spectral maps. Introduction InGaN/GaN hetero-structures are at the basis of many opto-electronic applications. For such applications both long and short range composition homogeneity is required. In this work we shall present the results of an extensive investigation on InGaN/GaN hetero-structures deposited by MOVPE on (0001) sapphire, using Ammonia, TMG and TMI precursors. The flow-dynamic conditions were carefully optimized in order to achieve a uniform layer thickness, but it was found that this condition is not sufficient to obtain homogeneus In incorporation. Composition profiles along the wafer radius were indeed detected by micro-Raman spectroscopy, spectrally resolved cathodo-luminescence (CL) and photoluminescence (PL). Generally the centre incorporates less indium than the periphery, however the composition profile is seen to change according to growth parameters (probably in connection with the MOVPE system flowdynamics). This information is a useful feedback for growth experiments. Furthermore, CL maps of the emission wavelength maximum suggests that local In clustering occurs. Experimental The growth of the InGaN layers was performed in a home-made MOVPE system using (0001) sapphire substrates and the standard Ammonia, TMG and TMI precursors. Typically the hetero-structure included a low-T GaN buffer (80-100 nm), a GaN layer deposited at high temperature (500-600 nm thick) and the InGaN layer deposited at 800-820 °C. In molar fraction was controlled either by changing the TMG/TMI flow ratio or by adding hydrogen into the reaction chamber. Micro-Raman measurements were carried out in backscattering geometry with a DILOR X-Y Micro-Raman spectrometer, using an Ar+ laser with line at 514.5 nm for excitation. Photoluminescence (PL) measurements were carried out both at room temperature and 12 K, using a He-Cd laser (line at 325 nm) for the excitation of the InGaN samples and a high-resolution monochromator for the analysis of the emitted light. Standard phototube and lock-in amplifier were used for signal detection. CL measurement were performed with an XiCLOne system from Gatan. The detection is done with a CCD camera, which allows a full spectral imaging, i.e. mapping of the spectral parameters of the selected luminescence band. The measurements were done at liquid nitrogen temperature (~80K). The spectra were typically obtained with an acceleration voltage of the e-beam of 5kV and beam current of 9nA, that give a spatial resolution around a few hundred nm. Results and discussion Figure 1a shows typical Raman spectra of InGaN alloys with different nominal In molar fractions, and a GaN reference spectrum which exhibits the typical allowed modes for hexagonal GaN in backscattering geometry: E 2 at 567 cm -1 and A 1 (LO) at 734 cm -1 . The InGaN layers present additional bands at about 522, 560 and 680 cm -1 . The line at 560 cm -1 is ascribed to the E 2 mode of the ternary alloy and is observed to shift linearly with the x molar fraction of In x Ga 1-x N, within the range 0 - 0.18, as reported in Figure 1b. The broad band around 680 cm -1 is seen to increase with the In content, suggesting that this Raman feature is probably generated by In 10th European Workshop on MOVPE, Lecce (Italy) 8-11 June 2003 PS.VI.13 R. Fornari et al