M. K. Horton

Segregation of in to dislocations in InGaN

Dislocations are one-dimensional topological defects that occur frequently in functional thin film materials and that are known to degrade the performance of InxGa1-xN-based optoelectronic devices. Here, we show that large local deviations in alloy composition and atomic structure are expected to occur in and around dislocation cores in InxGa(1-x)N alloy thin films. We present energy-dispersive X-ray spectroscopy data supporting this result. The methods presented here are also widely applicable for predicting composition fluctuations associated with strain fields in other inorganic functional material thin films.

Carrier localization in the vicinity of dislocations in InGaN

This project is funded in part by the European Research Council under the European Communitys Seventh Framework Programme (FP7/2007-2013)/ERC Grant Agreement No. 279361 (MACONS). The research leading to these results has received funding from the European Union Seventh Framework Programme under Grant Agreement 312483-ESTEEM2 (Integrated Infrastructure InitiativeI3). F.M. would also like to acknowledge the financial support from EPSRC Doctoral Prize Awards and Cambridge Philosophical Society. M.H. would like to acknowledge support from the Lindemann Fellowship.

Structure and electronic properties of mixed (a + c) dislocation cores in GaN

Classical atomistic models and atomic-resolution scanning transmission electron microscopy studies of GaN films reveal that mixed (a + c)-type dislocations have multiple different core structures, including a dissociated structure consisting of a planar fault on one of the {1 2¯10} planes terminated by two different partial dislocations. Density functional theory calculations show that all cores introduce localized states into the band gap, which affects device performance.

Dislocation core structures in Si-doped GaN

Aberration-corrected scanning transmission electron microscopy was used to investigate the core structures of threading dislocations in plan-view geometry of GaN films with a range of Si-doping levels and dislocation densities ranging between (5 ± 1) × 108 and (10 ± 1) × 109 cm−2. All a-type (edge) dislocation core structures in all samples formed 5/7-atom ring core structures, whereas all (a + c)-type (mixed) dislocations formed either double 5/6-atom, dissociated 7/4/8/4/9-atom, or dissociated 7/4/8/4/8/4/9-atom core structures. This shows that Si-doping does not affect threading dislocation core structures in GaN. However, electron beam damage at 300 keV produces 4-atom ring structures for (a + c)-type cores in Si-doped GaN.

Dislocations in AlGaN: Core Structure, Atom Segregation, and Optical Properties

We conducted a comprehensive investigation of dislocations in Al0.46Ga0.54N. Using aberration-corrected scanning transmission electron microscopy and energy dispersive X-ray spectroscopy, the atomic structure and atom distribution at the dislocation core have been examined. We report that the core configuration of dislocations in AlGaN is consistent with that of other materials in the III-Nitride system. However, we observed that the dissociation of mixed-type dislocations is impeded by alloying GaN with AlN, which is confirmed by our experimental observation of Ga and Al atom segregation in the tensile and compressive parts of the dislocations, respectively. Investigation of the optical properties of the dislocations shows that the atom segregation at dislocations has no significant effect on the intensity recorded by cathodoluminescence in the vicinity of the dislocations. These results are in contrast with the case of dislocations in In0.09Ga0.91N where segregation of In and Ga atoms also occurs but results in carrier localization limiting non-radiative recombination at the dislocation. This study therefore sheds light on why InGaN-based devices are generally more resilient to dislocations than their AlGaN-based counterparts.

Alloy fluctuations at dislocations in III-nitrides: identification and impact on optical properties

We investigated alloy fluctuations at dislocations in III-Nitride alloys (InGaN and AlGaN). We found that in both alloys, atom segregation (In segregation in InGaN and Ga segregation in AlGaN) occurs in the tensile part of dislocations with an edge component. In InGaN, In atom segregation leads to an enhanced formation of In-N chains and atomic condensates which act as carrier localization centers. This feature results in a bright spot at the position of the dislocation in the CL images, suggesting that non-radiative recombination at dislocations is impaired. On the other hand, Ga atom segregation at dislocations in AlGaN does not seem to noticeably affect the intensity recorded by CL at the dislocation. This study sheds light on why InGaN-based devices are more resilient to dislocations than AlGaN-based devices. An interesting approach to hinder non-radiative recombination at dislocations may therefore be to dope AlGaN with In.

Research data supporting "Dislocations in AlGaN: core structure, atom segregation and optical properties"

Figure 1. Plan-view aberration-corrected HAADF-STEM image of the AlGaN sample, showing the core structure of an edge-type dislocation (5/7-atom ring)(a,d,g), an undissociated mixed-type dislocation (double 5/6-atom ring)(b,e,h) and a dissociated mixed-type dislocation (7/4/8/4/9-atom ring)(c,f,i). Raw unfiltered images (a-c), and ABSF-filtered (average background subtraction filter) (d-f) with atomic columns identified to guide the eye (g-i). Figure 2. (a) Unfiltered HAADF-STEM image of an undissociated mixed-type dislocation. (b) ABSF-filtered image of (a) with geometric phase analysis overlay showing the x-x strain component (x-axis parallel to [11-20]). (c) EDX line scan showing the composition of Al, Ga, and N along the line depicted in (b) (with a ca. 1 nm analysis width). Figure 3. (a) AFM, (b) CL integrated intensity, and (c) CL peak emission energy of the same region in the AlGaN sample. Figure 4. (a) AFM, (b) CL integrated intensity, and (c) CL peak emission energy of the same region in the InGaN sample. Figure 5. Simulation of the emission energy shift in the vicinity of an edge-type dislocation.

Alloy composition fluctuations and percolation in semiconductor alloy quantum wells

Fluctuations in local alloy composition on small length scales may have a significant effect on device performance, particularly when there is a large disparity in the properties such as atomic size of the constituent alloy components. In particular, a random alloy is subject to a percolation threshold, above which an infinitely connected network of the minority alloy component exists. While these percolation thresholds are well known for ideal 2D and 3D lattices, they are unknown for the intermediary “2.5D” case, appropriate for quantum well structures. This letter presents calculations of the percolation threshold for 2.5D quantum well-like hexagonal, diamond/silicon and body-centred cubic lattices that are directly relevant to many semiconductor alloys, and enables further experimental inquiry into the effect of percolation on the properties of semiconductor alloys.

Research data supporting "Carrier Localization in the Vicinity of Dislocations in InGaN"

FIG. 1. AFM (a), SEM (b), panchromatic CL (c), and ADF-STEM (d) performed on the same micrometre-scale area. To guide the eye, a few dislocations are indicated by arrows in each picture. (e) High-resolution (HR) STEM image of the dislocation indicated by a square in (a)-(d), enabling the identification of the core structure (here dissociated 7/4/8/5-atom ring), and (f) geometric phase analysis (GPA) showing the e_xx strain component of the dislocation core region. FIG. 2. Schematic showing the electron probe in the SEM-CL scanning across a V-pit. The scale of the schematic, although indicative, is representative of the experimental conditions in which the experiment was conducted. Distance to nearest neighbor dependence of the intensity ratio (a)(c) and energy shift (b)(d) measured at the center (a)(b) and facet (c)(d) of the V-pits. FIG. 3. (a) Histogram of the number of In-N chains as a function of the number of indium atoms in the chains, located within a 10 A radius centered on the dislocation, in the case of a random distribution of indium (i.e. initial configuration of the simulation) or segregation of indium (i.e. equilibrium configuration of the simulation). Abstract representation of the data in (a), in the case of a random distribution (b) or segregation (c) of indium atoms. FIG. 4. ADF-STEM image of the clustered dislocations 26 (a) and 87 (b). The white strain-related contrast between the neighboring dislocations is indicated by an arrow. Aberration-corrected HAADF-STEM image of the core of dislocation 26 (dissociated 7/4/8/4/9-atom ring)(c) and 87 (undissociated double 5/6-atom ring)(d). An ABSF-filter (Average Background Subtraction Filter) has been applied to (c) and (d) in order to remove noise from the images. FIG. 5. 16K CL integrated intensity (a)(c) and peak emission energy (b)(d) maps of isolated (a)(b) and clustered (c)(d) dislocations. To guide the eye, the position of the bright spots, directly observable in (a) and (c), is indicated by circles in all the maps. To emphasize the relative variations in intensity and energy between isolated and clustered configurations, a common color scale is used in (a) and (c) and in (b) and (d).