P. G. Middleton

EXCITON LOCALIZATION AND THE STOKES' SHIFT IN INGAN EPILAYERS

We report a comparative study of the emission and absorption spectra of a range of commercial InGaN light-emitting diodes and high-quality epilayers. A working definition of the form of the absorption edge for alloys is proposed, which allows a unique definition of the Stokes’ shift. A linear dependence of the Stokes’ shift on the emission peak energy is then demonstrated for InGaN using experimental spectra of both diode and epilayer samples, supplemented by data from the literature. In addition, the broadening of the absorption edge is shown to increase as the emission peak energy decreases. These results are discussed in terms of the localization of excitons at highly indium-rich quantum dots within a phase-segregated alloy.

Optical linewidths of InGaN light emitting diodes and epilayers

A comparative study of the optical linewidths of photo- and electroluminescence from high-quality InGaN epilayers and commercial single quantum well light emitting diode structures was undertaken. Optical linewidths in both cases are temperature insensitive and increase systematically with increasing indium concentration. We assess the contribution of three mechanisms to the luminescence linewidth: alloy fluctuations, well width fluctuations, and strain effects. It is found that the broadening of the emission line is an intrinsic property of InGaN alloys. The piezoelectric effect in wurtzite semiconductor is proposed as a novel line-broadening mechanism.

The dependence of the optical energies on InGaN composition.

A wide-ranging experimental approach reveals a linear relationship between photoluminescence band peak energy and measured indium fraction for In[x]Ga[1] N epilayers with 0 < x < 0.40. We examine the dependence of the emission spectrum on composition using local measurements of the average indium content by Rutherford backscattering spectrometry, energy dispersive X-ray analysis, extended X-ray absorption fine structure and wavelength dispersed electron probe micro-analysis. Corresponding absorption and photoluminescence excitation data reveal the existence of a supplementary linear relationship between the optical bandgap and the indium fraction. Our observations provide definitive and conclusive evidence that the optical properties of InGaN do not conform to current theoretical descriptions of alloy band structure.

Photoluminescence of localized excitons in pulsed-laser-deposited GaN

Continuous-wave photoluminescence (PL) and time-resolved photoluminescence of gallium nitride layers grown by pulsed laser deposition are compared. The temperature dependence of the photoluminescence decay time and the PL-integrated intensity allows a determination of radiative and nonradiative time constants of GaN. We find that luminescence peaks centered at 3.360 and 3.305 eV at low temperature can be attributed to recombination of excitons localized at extended defects. The photoluminescence radiative lifetime at room temperature is on the order of tens of ns.

The optical and structural properties of InGaN epilayers with very high indium content

Abstract We present the results of optical and structural investigations of InGaN epilayers grown by Metallorganic Vapour Phase Epitaxy (MOVPE). The peak energies of characteristic photoluminescence (PL) bands allow us to identify regions of crystal with different mean InN: (InN+GaN) fraction in the range from 0.1 to nearly 1 in selected samples. The PL peak energy and the optical absorption band edge are strongly intercorrelated, the Stokes’ shift and the Urbach tailing energy both increase with InN fraction. High-resolution energy dispersive X-ray analysis (EDX), coupled with scanning electron microscopy (SEM) and cathodoluminescence (CL) imaging, helps to establish striking microscale correlations between optical and structural properties. Finally, X-ray absorption fine structure (XAFS) at the In and Ga K-edges reveals characteristic local structure on the atomic scale for InGaN solid solutions over the available range of In:Ga composition ratios.

Alternative substrates for gallium nitride epitaxy: photoluminescence and morphological investigations

Abstract Gallium nitride films grown by molecular beam epitaxy (MBE) on (0001) sapphire, lithium gallate and gallium arsenide (111)B substrates have been characterized using atomic force microscopy and photoluminescence spectroscopy. Inhomogeneities in the sapphire-based material are further explored via fluorescence imaging. Layers grown on GaAs substrate are shown to display superior luminescence properties and excellent surface morphology.

The Morphology and Cathodoluminescence of GaN Thin Films

In this paper we compare gallium nitride (GaN) films grown by molecular beam epitaxy on sapphire (Al 2 O 3 ), gallium arsenide (GaAs (111)B) and lithium gallate (LiGaO 2 ) substrates. Atomic force microscopy, scanning electron microscopy, cathodoluminescence imaging and cathodoluminescence spectroscopy are used to characterise the films. From growth runs carried out to date, GaN films on GaAs substrates exhibit the best surface uniformity and the cleanest luminescence.

Characterisation of nitride thin films by electron backscattered diffraction

Abstract In this paper, we describe the technique of electron backscattered diffraction (EBSD) and illustrate its use in the characterisation of nitride thin films by describing our results from a silicon-doped 3 μm thick GaN epilayer grown on a sapphire substrate misoriented by 10° towards the m-plane (10-10). We show that the EBSD technique may be used to reveal the relative orientation of an epilayer with respect to its substrate (a 90° rotation between the GaN epilayer and sapphire substrate is observed) and to determine its tilt (the GaN epilayer was found to be tilted by 12±3° towards [10-10] GaN ).

GaN epilayers on misoriented substrates

Abstract Three silicon-doped 3 μm thick GaN epilayers were grown simultaneously by metalorganic chemical vapour deposition on (0001) sapphire substrates misorientated by 0, 4 and 10° toward the m-plane (10 1 0). A comparative study of these epilayers was undertaken using photoluminescence (PL) spectroscopy, atomic force microscopy (AFM) scanning electron microscopy (SEM), cathodoluminescence (CL) imaging and CL spectroscopy. Low temperature PL of the 0 and 4° epilayers shows bound exciton (BE) emission between 3.47 and 3.48 eV and a low level of yellow band emission. The peak intensities of both emission bands are a factor of 2 higher for the 4° layer. In the 10° epilayer, the BE band is 3× stronger than in the 0o epilayer but there is no discernible yellow band. However, a number of additional bands appear at 3.459, 3.417, 3.362, 3.345, 3.309 and 3.285 eV. These bands may be attributed to the presence of structural defects in this epilayer, pointing to an abrupt degradation of its structural quality compared to the others. This degradation is confirmed by AFM studies. On a 20×20 μm2 image the 0 and 4° epilayers exhibit smooth surface morphologies, while the 10° epilayer shows a high density of hexagonal pits. Finally, SEM images reveal the surface of the 10° epilayer to be ‘streaked’ and pitted. Low temperature CL images at 3.48 eV (bound exciton region) show random spotty emission, while those at 3.28 eV and 3.41 eV exhibit a streaky appearance similar to the SEM image. This suggests that these luminescence bands are indeed associated with structural defects.

Energy-dispersive x-ray imaging of an InGaN/GaN bilayer on sapphire.

In this letter we discuss the potential and the limitations of quantitative characterization of the distribution of In and Ga atoms in III–N mixed alloys using energy dispersive x-ray (EDX) analysis. Spatial fluctuations of the indium content in an InGaN epilayer are found to correspond to changes in luminescence efficiency. Large hexagonal pyramids, which appear sparsely in such layers, appear to be relatively deficient in indium. Monte Carlo simulations, used to profile the Ga Ka x-ray fluorescence, highlight several limitations of the EDX technique in this context, but confirm our interpretation of the data. Finally, we identify differential growth rates as a possible explanation for the concentration/efficiency variations in InGaN layers.