Influence of composition, strain, and electric field anisotropy on different emission colors and recombination dynamics from InGaN nanodisks in pencil-like GaN nanowires

Ž. Gačević, N. Vukmirović, N. García-Lepetit, A. Torres-Pardo, M. Müller, S. Metzner, S. Albert, A. Bengoechea-Encabo, F. Bertram, P. Veit, J. Christen, J. M. González-Calbet, and E. Calleja

Phys. Rev. B 93, 125436 – Published 30 March 2016

Abstract

This work reports an experimental and theoretical insight into phenomena of two-color emission and different electron-hole recombination dynamics in InGaN nanodisks, incorporated into pencil-like GaN nanowires. The studied nanodisks consist of one polar (on $c$ facet) and six (nominally) identical semipolar (on $r$ facets) sections, as confirmed by transmission electron microscopy. The combination of cathodoluminescence with scanning electron microscopy spatially resolves the nanodisk two-color emission, the low-energy emission (∼500 nm) originating from the polar section, and the high-energy emission (∼400 nm) originating from the semipolar section. This result has been directly linked to a “facet-dependent” nanodisk composition, the In content being significantly higher in the polar ( $\sim 20 %$ ) vs semipolar ( $\sim 10 %$ ) section (as quantified by energy dispersive x-ray spectroscopy), further leading to a strong facet-dependent strain anisotropy. Time-resolved cathodoluminescence reveals significantly different electron-hole recombination times in the two sections, moderately fast (∼1.3 ns) vs fast (∼0.5 ns) in polar/semipolar sections, respectively, the difference being linked to a strong anisotropy in the nanodisk internal electric fields. To determine the influence of each of the three contributing “facet-related” anisotropies (composition, strain, and electric field) on the two-color emission, a proper simulation [relying on virtual crystal approximation and involving three-dimensional (3D) continuum mechanical modeling, a 3D Poisson equation, and a one-dimensional Schrödinger equation] has been performed. The theoretical simulations allow the three effects to be quantitatively disentangled, revealing a clear hierarchy among their contributing weights, the facet-dependent composition inhomogeneity being identified as the dominant one (and the strain inhomogeneity being identified as the least significant one). As for different recombination times, while it is mainly linked to the internal electric field anisotropy, we also suggest that it is, very likely, influenced by gradually increasing In content along the nanodisk growth direction (lattice-pulling effect); the latter mechanism keeps electrons and holes in (relative) proximity within the polar section, enabling their relatively fast and efficient radiative recombination.

Received 12 November 2015
Revised 26 January 2016

DOI:https://doi.org/10.1103/PhysRevB.93.125436

Physics Subject Headings (PhySH)

Electronic structure

Nanowires

Energy-dispersive x-ray spectroscopy k dot p method

Condensed Matter, Materials & Applied Physics

Authors & Affiliations

Ž. Gačević^1,*, N. Vukmirović², N. García-Lepetit¹, A. Torres-Pardo^3,4, M. Müller⁵, S. Metzner⁵, S. Albert¹, A. Bengoechea-Encabo¹, F. Bertram⁵, P. Veit⁵, J. Christen⁵, J. M. González-Calbet³, and E. Calleja¹

¹ISOM-ETSIT, Universidad Politécnica de Madrid, Avenida Complutense s/n, 28040 Madrid, Spain
²Scientific Computing Laboratory, Institute of Physics Belgrade, University of Belgrade, Pregrevica 118, 11080 Belgrade, Serbia
³Departamento de Química Inorgánica, Universidad Complutense, 28040, Madrid, Spain
⁴CEI Campus Moncloa, UCM-UPM, Madrid, Spain
⁵Institute of Experimental Physics, Otto-von-Guericke-University Magdeburg, 39106 Magdeburg, Germany

^*gacevic@isom.upm.es

Article Text (Subscription Required)

Click to Expand

Supplemental Material (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand

Issue

Vol. 93, Iss. 12 — 15 March 2016

Reuse & Permissions

Access Options

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
(a) Top SEM view of a representative sample area revealing ordered hexagonal array of pencil-like NWs. (b) TEM micrograph of a representative single NW (sample B) revealing the structure of its building blocks. (c) Atomic resolution HAADF STEM reveals structural properties of the GaN-capping/InGaN-nanodisk interface: (up) polar c-facet and (down) semipolar s-facet interface.
Reuse & Permissions
Figure 2
(a) A NW probing is performed along the central axis (red line) with the $e$ beam penetrating perpendicularly to it (green arrow). Different probing heights (I, II, and III) have been used to assess different building blocks: (I) GaN-capping/semipolar-InGaN/core-GaN/semipolar-InGaN/GaN-capping, (II) GaN-capping/polar-InGaN/GaN-capping, and (III) GaN-capping only. (b) Spatial distribution of the In- $L α_{1}$ line intensity yields information about In incorporation throughout the NW tip region. (c) EDS spectra taken at three heights, I, II, and III, confirm In presence in the InGaN nanodisk, exclusively, its content being significantly higher in the polar ( $\sim 20 %$ ) vs semipolar ( $\sim 10 %$ ) section.
Reuse & Permissions
Figure 3
(a) SEM image combined with monochromatic room temperature CL measurements, performed at 405 nm (violet) and 505 nm (green) wavelengths. (b) ADF STEM image of a single NW combined with a highly spatially resolved CL mapping (at 15 K). The images reveal that the green luminescence originates from the top-facet InGaN, and violet luminescence originates from side-facet InGaN, whereas the UV luminescence comes from the core GaN (the wavelength is color coded).
Reuse & Permissions
Figure 4
(a) Low-temperature PL (10 K) spectra show three intense peaks attributed to GaN, semipolar InGaN, and polar InGaN, respectively. (b) Power-dependent PL measurement shows that when the PL excitation intensity is increased two orders of magnitude, only the polar-InGaN emission blueshifts (around 30 meV). (c) TR CL measurements (∼5 K) reveal monoexponential and nonmonoexponential CL decays for semipolar and polar InGaN, respectively. (d) Temperature-dependent TR CL reveals significantly shorter initial decay time of semipolar vs polar InGaN. In addition, the initial decay times of the polar InGaN sections are strongly affected by the nanodisk thickness, being shortest in the case of sample A.
Reuse & Permissions
Figure 5
Theoretical simulations for sample B. (a) Hydrostatic strain $ɛ_{x x} + ɛ_{y y} + ɛ_{z z}$ (left) at the NW cross section and (right) along the NW central axis (green arrow). (b) Electric field intensity (left) at the NW cross section and (right) along the NW central axis (orange arrow). (c) Conduction and valence band edge energies at the NW cross section. (d) Band-gap profiles in the semipolar and polar nanodisk sections, along the trajectories designated in (c).
Reuse & Permissions
Figure 6
CB and VB profiles along the direction shown in Fig. 5. (Solid line) pulling effect considered, assuming linearly graded InGaN in the polar section: $X_{In} = 0 % - 20 %$ . (Dotted line) pulling effect not considered, assuming uniform InGaN composition in the polar section: $X_{In} = 20 %$ .
Reuse & Permissions

Physical Review B

covering condensed matter and materials physics