Nonuniform indium distribution within InGaN/GaN single quantum well (SQW) structures with nanoscale islands grown by metalorganic chemical vapor deposition (MOCVD) have been characterized by advanced characterization techniques. Robinson backscattered electron (BSE) measurements show cluster‐like BSE contrast of high brightness regions, which are not centered at small dark pits in a SQW structure of spiral growth mode. By comparing with the secondary electron (SE) images, the bright cluster areas from the BSE images were found to have higher indium content compared to the surrounding dark areas. Temperature dependant photoluminescence (PL) measurement shows typical “S‐shape” curve, which shows good correlation with nonuniform indium distribution from BSE measurement. Optical evaluation of the samples show increased PL slope efficiency of the spiral mode SQW, which can be attributed to the presence of Indium inhomogeneities. (© 2005 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)

This study exemplifies the use of TappingMode TM atomic force microscopy (AFM) surface morphology imaging to investigate and optimise the metalorganic chemical vapour deposition (MOCVD) growth conditions and post-growth stability of thin (<40 Å) InGaN layers with direct implications to the structural and optical properties of blue (460 nm) and green (520 nm) LEDs. InGaN epilayers less than 40 Å thick of ~ 20% solid phase indium were produced on thick (3-4 µm) 2" GaN templates grown on (0001) c-plane sapphire substrates. The morphological evolution of the InGaN material was studied utilising a DI3100 AFM tool. Surface morphology and its correlation with photoluminescence and X-ray diffraction results will be discussed for every set of conditions employed. More specifically, the post-growth ambient exposure and thermal stability of the uncapped InGaN epilayers were investigated. In addition, the initial stage of subsequent GaN growth, which is an essential step towards the manufacture of LED active regions, was examined. Based on the above findings, a flexible MOCVD growth parameter space and improved LED constituent layer sequencing techniques have been established leading to more efficient and stable LED devices.