Analysis of InGaN/GaN single quantum wells by X‐ray scattering and transmission electron microscopy

Abstract

A set of In x Ga 1-x N/GaN single quantum wells (SQWs) were analysed using X-ray reflectivity (XRR), high resolution X-ray diffraction (HRXRD) and high resolution transmission electron microscopy (HRTEM). A method for determining the thickness and composition of the single quantum well layers using only X-ray scattering measurements is described -the results agree very closely with those determined from HRTEM analysis of the same samples. HRTEM lattice fringe images obtained from InGaN subjected to minimal exposure to the electron beam were used to determine local lattice parameter maps. These show the indium to be distributed uniformly within the alloy. The electron beam very rapidly caused damage to the InGaN which could be falsely interpreted as an inhomogeneous indium distribution.1 Introduction Accurate characterisation of In x Ga 1-x N quantum well (QW) widths, compositions and alloy homogeneity is an important stage in understanding the growth and optical properties of InGaN based LEDs and laser diodes. Though InGaN QW widths can be routinely measured using HRTEM, in many situations X-ray scattering measurements are advantageous over TEM experiments: they are nondestructive, average over a greater amount of material and can yield information more rapidly. Characterisation of QW structures using HRXRD relies on optimising the parameters in a model so that its simulated scattering (based on X-ray scattering theory) most closely matches the measured diffraction behaviour. For InGaN/GaN QWs lying normal to [0001] this is best achieved using an ω/2θ scan about the GaN (0002) reflection [1]. Care must be taken when using this approach since different parameters in the model can result in indistinguishable scattering simulations for scattering angles close to the primary GaN reflection. For multiple quantum well (MQW) structures it is necessary to consider the intensities of several orders of the superlattice satellite diffraction peaks to distinguish between different models which generate otherwise similar scattering [2]. This approach has recently been used to extract QW thicknesses and compositions from InGaN/GaN MQW structures with five and ten superlattice repeats [1]. The GaN (0002) ω/2θ scan from an InGaN SQW structure (Fig. 1) does not contain any superlattice satellite peaks (only thickness fringes) and the information which can be obtained in a reasonable time using a laboratory diffractometer cannot be used on its own to ascertain the structure. In this paper we describe how complementary XRR measurements were used together with the HRXRD data to determine layer thicknesses and compositions for a set of InGaN/GaN SQW structures. HRTEM analysis indicates the parameters measured by X-ray scattering to be reliable and that the InGaN alloy is homogeneous in these samples. The potential for detecting false indium clustering due to electron beam damage of the InGaN is also identified.The wurtzitic InGaN/GaN heterostructures were grown on c-plane sapphire substrates by metalorganic vapour phase epitaxy (MOVP...

Does In form In-rich clusters in InGaN quantum wells?

“…At low electron dose, the lattice fringe image of the quantum well and the lattice parameter map are both reasonably uniform ( figure 1a, 1c), [16,17]. We have studied the effect of 200, 300, and 400 keV incident electrons.…”

“…In the light of the Smeeton et al [16,17] papers, which suggested that the gross indium-rich clusters in InGaN quantum wells reported by many researchers might be due to electron beam damage, the Gerthsen group revised their earlier conclusions [11,12]. They observed that the indium concentration in the clusters increased with increasing irradiation time in the electron microscope.…”

“…Insert Fig.1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Since publishing the Smeeton et al [16,17] papers it has been suggested to us that our results may apply only to InGaN grown on MOVPE equipment at Cambridge, or may be related to our TEM specimen preparation procedures, rather than being a general effect. We have therefore purchased a very bright blue commercial LED and examined the InGaN quantum wells it contains by TEM.…”

“…We will now demonstrate that in the wide range of InGaN materials we have examined, such gross indium-rich clusters do not exist, and they are produced by electron beam damage in the TEM. resolution TEM, and the compositions measured using X-ray reflectivity and high resolution X-ray diffraction as described elsewhere [15,16].…”

“…Figure 1a was recorded within 20 seconds of first exposing this part of the quantum well to the electron beam, figure 1b is the same area after a few minutes of exposure. We have analysed these images to produce lattice parameter maps [16,17] using a process similar to the DALI technique [11]. After only a few minutes exposure to the electron beam we found nanometre-size indium clusters formed which caused local strains of up to 10%, corresponding to an indium fraction x of 60%.…”