Nanoscale Imaging of InN Segregation and ... - ACS Publications

Letter pubs.acs.org/NanoLett

Nanoscale Imaging of InN Segregation and Polymorphism in Single Vertically Aligned InGaN/GaN Multi Quantum Well Nanorods by TipEnhanced Raman Scattering E. Poliani,*,† M. R. Wagner,†,‡ J. S. Reparaz,†,‡ M. Mandl,§ M. Strassburg,§ X. Kong,∥ A. Trampert,∥ C. M. Sotomayor Torres,‡,⊥ A. Hoffmann,† and J. Maultzsch† †

Institut für Festkörperphysik, Technische Universität Berlin, 10623 Berlin, Germany ICN2Institut Catala de Nanociencia i Nanotecnologia, Campus UAB, 08193 Bellaterra (Barcelona), Spain § OSRAM Opto Semiconductors GmbH, 93055 Regensburg, Germany ∥ Paul Drude Institute, 10117 Berlin, Germany ⊥ Institució Catalana de Recerca I Estudis Avançats (ICREA), 08010 Barcelona, Spain ‡

S Supporting Information *

ABSTRACT: Vertically aligned GaN nanorod arrays with nonpolar InGaN/ GaN multi quantum wells (MQW) were grown by MOVPE on c-plane GaNon-sapphire templates. The chemical and structural properties of single nanorods are optically investigated with a spatial resolution beyond the diffraction limit using tip-enhanced Raman spectroscopy (TERS). This enables the local mapping of variations in the chemical composition, charge distribution, and strain in the MQW region of the nanorods. Nanoscale fluctuations of the In content in the InGaN layer of a few percent can be identified and visualized with a lateral resolution below 35 nm. We obtain evidence for the presence of indium clustering and the formation of cubic inclusions in the wurtzite matrix near the QW layers. These results are directly confirmed by high-resolution TEM images, revealing the presence of stacking faults and different polymorphs close to the surface near the MQW region. The combination of TERS and HRTEM demonstrates the potential of this nanoscale near-field imaging technique, establishing TERS as a very potent, comprehensive, and nondestructive tool for the characterization and optimization of technologically relevant semiconductor nanostructures. KEYWORDS: Tip-enhanced Raman scattering, optical near-field spectroscopy, subwavelength chemical imaging, InGaN quantum wells, indium segregation

T

quantum efficiency due to the quantum confined Stark effect, can both be avoided by InGaN/GaN nanorods with nonpolar QWs.5 Despite some achievements, the efficiency droop for high In concentrations remains a central challenge for future lightemitting devices. The key toward the improvement of the luminance efficacy in InGaN structures must therefore be found in the reduction of nonradiative recombination centers such as defects/dislocations,6−9 surfaces/interfaces,10 and mechanisms such as Auger recombination,11 polarization fields,12 carrier leakage/overflow,13,14 and segregation/clustering of indium.15,16 Consequently, detailed information about the extent, dimension, and quantity of the different nonradiative recombination centers at the nanoscale is highly desirable to optimize the next generation of light-emitting devices.

he rapid progress in the growth and fabrication of semiconductor nanostructures in recent years continues to attract extensive research efforts toward both monitoring and tailoring of material properties at the nanoscale. As an example, InGaN based nanostructures constitute a promising material class for the next generation of highly efficient, large-scale solid state lighting devices.1 The wide bandgap tunability of InGaN has established this ternary system as desirable material for colored light-emitting diodes (LEDs) and laser diodes over the entire range of the visible spectrum.2 An inherent limitation of the current nitride-based technology is the large lattice mismatch to the commonly used sapphire substrates and the presence of strong piezo- and pyroelectric fields. Compared to conventional thin film structures, nanorods and nanowires release strain at the interface to sapphire substrates.3 In addition, InGaN/GaN multi quantum well growth was reported on nonpolar faces of GaN nanowires.4 Thus, two main limitations for the device performance, namely, the high dislocation density due to large lattice mismatch and the low © 2013 American Chemical Society

Received: April 10, 2013 Revised: June 17, 2013 Published: June 24, 2013 3205

dx.doi.org/10.1021/nl401277y | Nano Lett. 2013, 13, 3205−3212

Nano Letters

Letter

Figure 1. Schematic illustration of the TERS setup and sample structure. The morphology of the sample is obtained by scanning tunneling microscopy (STM). The STM nanopositioning system is used to acquire the tip-enhanced Raman map on the top of a single nanorod. The STM current feedback assures a constant distance between the sample surface and the STM tip during the measurement.

with a spatial resolution of about 35 nm.24−26 We identify cubic and hexagonal InN modifications near the InGaN MQW region solely based on the TERS data which is confirmed independently by HRTEM images. In addition, local fluctuations in the In concentration can be visualized. Furthermore, we report the first observation of narrow surface optical Raman modes of a single GaN nanorod. The nanorod sample was grown in a large-volume metal− organic vapor phase epitaxy (MOVPE) reactor on a GaN-onsapphire template. A 30 nm SiO2 layer was deposited after template growth and subsequently patterned by a commercial nanoimprint lithography process. After etching of the resist, regular holes with diameters of 260 nm in the SiO2 mask were obtained. The nanorods were nucleated in these openings, applying growth conditions which support 3D growth as described in ref 27. Subsequently to the deposition of the 1020 cm−3 silicon-doped GaN core, an In0.2Ga0.8N/GaN double quantum well and a 3 nm thick GaN cap layer were grown in 2D growth mode. The geometrical properties of the nanorods (nearly constant diameter of 300 nm at an aspect ratio of 5) were investigated by scanning electron microscopy (SEM), revealing only small variations in the dimensions of the nanorods with a homogeneous shape throughout the entire 4” wafer. So-called pit-like crystalline defects were found at the pyramidal shaped tip of the nanorods due to the relatively low temperatures during the shell growth.

On the micrometer scale, information about composition, strain, and structural quality can be easily obtained by a variety of techniques including micro-Raman spectroscopy. However, the optical diffraction limit renders this approach not applicable at the nanoscale. Nanoscale materials properties are therefore only accessible by local and partly destructive techniques such as high-resolution transmission electron microscopy (HRTEM) imaging, which underlines the need for an alternative, nondestructive technique. A possibility to circumvent the spatial resolution limit for optical measurements is given by optical near-field methods such as tip-enhanced spectroscopy (TERS).17−23 In this Letter, we demonstrate the potential of tip-enhanced Raman scattering to obtain a variety of materials properties in technologically relevant InGaN semiconductor nanostructures with a spatial resolution well below the diffraction limit. TERS combines the morphological characterization of a scanning tunneling microscope (STM) with the optical characterization by micro-Raman spectroscopy. Moreover, the enhanced nearfield sensitivity of this method is capable of providing detailed information about structures with extremely low scattering cross sections such as interfaces and surfaces as well as local variations of the carrier concentration. By analyzing TERS maps of a single InGaN MQW nanorod, we obtain a multitude of information on the chemical composition, crystallographic modifications, local strain variations, and charge accumulations 3206


Nano Letters

Letter

dimensions of the STM tip and its piezo-positioning capabilities and thus is expected to be even higher (approximately 10−20 nm), as supported by the abrupt changes in the Raman spectra between two adjacent points. The acquisition time for each Raman spectrum was limited to 15 s to avoid artifacts due to thermal drift. The peak positions of the different Raman modes were obtained by numerical fitting of each spectrum. As an example, Figure 3b shows a color-coded intensity map for the strongest lines in the Raman spectra between 457 and 480 cm−1 (c.f. Figure 4a,b). The scanning electron microscopy (SEM) image of a typical nanorod is displayed in Figure 3c. By combining STM and TERS data, one can obtain a 3D visualization of the surface topology correlated to, for example, local variations in the crystal structure as discussed below (Figure 3d). Figure 4a shows the far-field micro-Raman spectrum (retracted STM tip) and two tip-enhanced Raman spectra at spatial positions close to the apex of the nanorod in comparison to a bulk GaN reference spectrum. The far-field spectrum is dominated by the E2(high) mode of GaN at 568 cm−1, which originates from the core region of the nanorod, whereas the additional TERS modes from the near surface volume are not observed in the far field (c.f. Figure 4a). This position coincides with the unstrained value of the E2(high) mode in bulk GaN as seen in the reference spectrum.32 Additional Raman modes, which are undetectable in the far-field spectrum, appear when the gold tip is approached to the sample because of the nearfield interaction. We estimate a factor of 250 as lower limit for the enhancement in the near-field spectra by background subtraction and normalization of all spectra to the peak intensity of the E2(high) mode, such that the intensity of the tip-enhanced modes would be comparable to the noise level of the far-field spectrum. Apart for the omnipresent E2(high) mode, three features in the tip-enhanced Raman spectra are of particular interest as they reveal a clear trend throughout the entire Raman map. A representative example is given in Figure 4b, which displays the spectra for a line scan near the center of the nanorod. The first peaks are observed between 457 and 480 cm−1, the second ones between 518 and 533 cm−1, and the third ones between 640 and 650 cm−1. Although the frequencies of these modes vary over the spatial map, they do not drift or vary randomly. Instead, the peaks appear at few distinct spectral positions throughout the map. This excludes artifacts such as thermal effects or tip contamination. The frequencies of these Raman modes are plotted in Figure 5a−c as function of their spatial coordinates in a 3D plot and in Figure 5e−g for the spectra of the line scan in Figure 4b. We show below how these data reveal information about variation of strain, composition, segregation, and crystal structure on the nanoscale. We start by analyzing the shift of the nonpolar, that is, exclusively strain sensitive E2(high) mode. The frequency of this mode is 568 cm−1 and remains constant within the spectral resolution over the entire map. Two possible origins for this mode are conceivable: (i) a far-field signal from the GaN nanorod core region as shown in Figure 1, or (ii) a tipenhanced signal originating from the GaN cap layer of the rods. In the latter case, a dominating A1(TO) from the GaN cap layer around 533 cm−1 would be expected33 since the A1(TO) mode was shown to exhibit a stronger TERS enhancement than the E2(high) mode in GaN thin films.20 Indeed, we observe a mode which strongly varies in frequency between 518 and 533 cm−1 for different spatial positions (see Figure 5a). However, the fact

The morphological characterization of the nanorods with subnanometer precision was achieved using a scanning tunneling microscope (PARK XE-100). The gold tips were produced by electrochemical etching following the common procedure described in refs 28 and 29. TERS measurements were performed by coupling a confocal Raman instrument (Horiba-Jobin Yvon LabRam HR-800) with 1 cm−1 spectral resolution to the STM, using a 50× long-distance microscope objective (numerical aperture NA = 0.45) tilted by 30° degrees with respect to the sample plane (c.f. Figure 1). A He−Ne laser at 632.8 nm was focused on the plasmonic optical cavity which is created by the gold tip of the STM and the sample surface, thus resonantly exciting the localized surface plasmons (LSPs) of the tip. The electric field of the incident laser light is locally enhanced at the tip by the lighting rod effect and the interaction between LSPs and surface plasmon polaritons (SPP) of the doped specimen when excited in resonance with the laser.30,31 As a consequence, the spatial resolution limit can in principle be reduced to about 15 nm or less.22 The strong localization of the evanescent field leads to a restriction of the sampling volume close to the surface of the nanorod. Considering the total thickness of the capping and QW layers as well as the HRTEM images, we approximate the vertical extent of the sampling volume to about 10 nm in accordance with reported literature values.17 High-resolution transmission electron microscopy (HRTEM) studies near the tip of individual nanorods reveal a nonuniform growth of the InGaN QWs as shown in Figure 2a. The HRTEM image in Figure 2b shows different localized

Figure 2. (a) Cross-section HRTEM image of the top of a single vertically aligned GaN/InGaN nanorod. The inset shows a lowresolution TEM image; the area of the HRTEM image is indicated by the square. (b) Cubic inclusions (ZB) and stacking faults (SFs) in the regular wurtzite (WZ) crystal structure are visible beneath the surface of the nanorod.

regions which crystallize in the wurtzite and zinc-blende structure. The cubic phase mainly occurs in the top area of the nanorods 5−10 nm below the apex of the rods, that is, where the QWs are expected according to the growth procedure. Apparently, the cubic regions in the QW layers induce stacking faults which appear in the GaN cap layer near the surface of the nanorods. To obtain Raman maps with high spatial resolution, the morphology of a single nanorod was measured by STM and subsequently mapped by TERS. An STM image of a nanorod overlaid by the coordinate system for the corresponding Raman map is displayed in Figure 3a. The surface area of 285 × 285 nm was mapped by 8 × 8 points, resulting in a distance between the points of two adjacent Raman spectra of 35 nm. The intrinsic spatial resolution is only limited by the apex 3207


Nano Letters

Letter

Figure 3. (a) AFM image of the top of a single nanorod. The surface is divided into a grid of 64 squares superimposed to the AFM image. The squares of the grid are identified by row and column numbers (r,c) which label the TERS spectra acquired at the center of each square. The size of the squares (pixel) represents the image resolution. (b) Intensity map of the TERS peaks for the cubic and hexagonal polymorphs of InN in the same coordinate system as shown in a. The green line marks the interface between the two different modifications. (c) SEM image of a single nanorod of the same sample. (d) Color-coded Raman intensity of the two different InN modes as texture of the three-dimensional morphology image acquired with the STM. Red indicates cubic InN clusters, blue hexagonal ones.

therefore reflect the presence of cubic and hexagonal InN clusters (denoted as c-InN and h-InN) near the top of the nanorods (see Table 1). The direct confirmation for the existence of cubic and wurtzite structures a few nanometers below the nanorod apex is given by the HRTEM images in Figure 2. Figure 4b shows the TERS spectra for a line scan along the fourth column of the coordinate system in Figure 3a. While the spectra (3,4) and (4,4) reveal the exclusive presence of the h-InN mode, the rest of the line scan shows only the cInN mode (Figure 5f). The same pattern systematically occurs for the other columns as shown in the Raman shift map in Figure 5b. This step-like behavior in the phonon frequency map divides the top of the nanorod into two distinct regions which are dominated by the presence of two different InN polymorphs formed by InN clusters due to In segregation, that is, a dense region of InN islands embedded into the InGaN layers.37,38 A vertical interface dividing the domains with the two polytypes is formed during the nucleation of the QW layers

that the E2(high) and the A1(TO) modes of GaN have similar phonon deformation potentials34 excludes the possibility of a common origin of these two lines. The absence of a pure GaN A1(TO) mode proves that the observed E2(high) mode of GaN cannot be a near-field enhanced mode but must originate from the GaN core region of the nanorod. Consequently, the A1(TO) mode must stem from the underlying InGaN QW layers. The shift of this mode thus reflects the combined effects of local variations in the In content of the ternary InGaN alloy and the presence of different local strains as discussed below. The spectra in Figure 4a are dominated by intense modes at 457 cm−1 (red spectrum) and 473 cm−1 (blue spectrum), which appear mutually exclusive depending on the spatial position of the TERS measurement. The frequencies of these modes are in excellent agreement with the theoretical and experimental values of the pure InN TO phonons in the cubic zinc-blende structure and the InN E1(TO) phonons in the hexagonal wurtzite modification, respectively.35,36 The observed modes 3208


Nano Letters

Letter

Figure 4. (a) GaN c-plane Raman reference spectrum (bottom), GaN/InGaN nanorods far-field Raman spectrum (black), and local TERS spectra at coordinates (5,5) (blue) and (4,5) (red). Spectra are normalized to the E2(high) peak maximum amplitude. (b) Raman spectra for a line scan along the fourth column of the TERS map. The fitted peak positions are plotted in Figure 5.

spatial distribution of the mode shift. Neglecting the strain contribution, the average wavenumber of 527 cm−1 for the A1(TO) mode does not exactly match the expected value for the In0.2Ga0.8N A1(TO) phonon but rather corresponds to an indium concentration of about 7% in the InGaN layer.43,44 During the segregation process, indium aggregates in clusters which depletes the surrounding area. This process is reflected by the lower than expected value for the measured In content. The fluctuations of the Raman shift around the average value indicate local variations of the indium concentration between 5% and 10% based on experimental studies of the A1(TO) phonon frequencies as a function of the indium content in InGaN.44,45 The spatial randomness of these variations and the match of the InN phonon frequencies with the bulk values support the assumption that strain does not play a significant role. The maximum value of 533 cm−1 at the center of the map demonstrates the presence of a small indium free region of pure hexagonal GaN33 at the apex of the nanorods. The extent of this pure GaN region is less than 35 nm. Figure 5a also reveals the presence of a strongly localized valley in the map of the InGaN A1(TO) mode which width is comparable with the resolution of the TERS map, that is, about 35 nm. In this region, the phonon frequency exhibits the lowest values of about 520 cm−1. The valley is located at the cubic nucleating side of the planar vertical defect as suggested in the model of Figure 5d and illustrated in the TEM image in the Supporting Information. Considering the disordered nature of the interface between the two polytypes, the surrounding InGaN layer can be subject to local strain. The induced strain propagates vertically inside the InGaN layer above the c-InN clusters. We therefore interpret the sharp valley in this frequency map as a highly localized region where the frequency shift is dominated by tensile strain rather than by fluctuations of the In content. A rough estimation of the residual stress in the valley region amounts to an average value of 0.7 GPa46 based on the Raman shift as compared to the adjacent points of around 3 cm−1 and the determined low average In content in the InGaN layer. This demonstrates that TERS is an efficient tool to determine strain and indium distribution on the nanoscale. The relationship between local sample properties

along the growth direction. The interface is located in the middle of these domains as indicated by the green line in the Raman intensity map of the InN modes in Figure 3b. The nucleation of InN clusters in this biphasic region could be induced by a preexistent inclusion of cubic GaN in the nanorod core or a planar vertical defect which reaches up to the surface of the nanorod. During the subsequent deposition of the QW layers, the exposed cubic crystallographic orientation of GaN on the surface nucleates the cubic modification of InN clusters only from one side of the interface. The growth of wurtzite InN clusters continues instead on the other side, propagating the underlying interface between hexagonal and cubic modification. The dimension of the interface between the regions containing clusters of h-InN and c-InN must be smaller than the lateral image resolution of 35 nm, since no coexistence of both InN modes is observed in any of the spectra. A schematic illustration of this model is shown in Figure 5d. The large enhancement of the TERS intensity for these modes can be understood considering the small bandgap of c-InN (Eg = 0.56 eV)39 and hInN (Eg = 0.68 eV)40 as compared to the surrounding InGaN. The difference causes a large carrier concentration in the InN clusters which leads to a strong coupling of SPPs and LSPs. Furthermore, the carrier concentration is expected to be additionally increased by charge accumulation close to the interfaces,41 which leads to an increased plasmon population and consequently a stronger enhancement of local InN modes. In case of the A1(TO) mode of the InGaN QW layers, the 3D map of phonon frequencies vs spatial coordinates and the Raman shifts for the line scan along the fourth column (r,4) are shown in Figure 5a and e, respectively. A behavior strictly related to the step-like character of the InN mode is observed for this mode, supporting our previously introduced model (Figure 5d). However, in contrast to the InN mode, the spectral position of the A1(TO) mode arising from the InGaN QW does not only depend on strain but also on the local fluctuations of the In content in the ternary system. Thus, it is not straightforward to distinguish between these two contributions to the mode frequency.42 Nevertheless, an estimation of the local fluctuations of the In content within the InGaN layers is still possible by a careful analysis of the 3209


Nano Letters

Letter

Figure 5. (a−c) Three-dimensional visualization of the Raman frequencies of three different modes as a function of the spatial map coordinates (r,c). (a) A1(TO) mode of InGaN layers, (b) E1(TO) and TO modes of the hexagonal and cubic modifications of InN, respectively, and (c) A1 surface optical (SO) mode of the GaN capping layer. Spatial coordinates without Raman signal of the respective modes are omitted from the maps. The arrows indicate the plotted column (fourth column) in the line scan profiles in e−g. (d) Simplified schematic illustration of the nanorod top region in cross sectional view. (e−g) Raman shift related to the cross sectional scheme in d for the spectra of the line scan in Figure 4b.

precondition for sensing the surface optical phonons is a perturbation of the surface potential which results in the relaxation of the scattering selection rules. The SEM and STM images in Figure 3c and d indeed show large topographical fluctuations which suggest a large surface roughness at the top of the nanorods. Earlier studies have revealed that the frequency and line width of the surface optical modes depend on the density and diameter of the nanorods.49,50 However, the

such as strain, composition, and polymorphism and the different TERS modes as well as HRTEM information are summarized in Table 1. Finally, we turn to the remaining group of Raman modes observed between 640 cm−1 and 650 cm−1. The frequencies of the modes in this range coincide with the expected values for surface optical phonons SO(A1) of the GaN cap layer which were frequently reported in GaN based nanostructures.47−49 A 3210


Nano Letters

Letter

scattering maps of single vertically aligned GaN nanorods with InGaN/GaN multiquantum wells. These properties include information on nanoscale variations of the chemical composition, indium clustering, crystallographic modification of clusters, local strain, and charge accumulation. The presence of compositional fluctuations and indium segregation inside the quantum wells was resolved with a spatial resolution of 35 nm. The abrupt change of the InN TO phonon frequency in the QW layers revealed the coexistence of hexagonal and cubic modifications in indium-rich regions, as confirmed by HRTEM. These regions were interpreted as biphasic zones composed of an agglomerate of InN islands which nucleate on top of a cubic inclusion in the GaN core. The presence of local strain in the region surrounding the interface between the two cluster polytypes is expressed by a local minimum in the energy of the A1(TO) InGaN phonons. With increasing distance from the interface, the spatially resolved changes in the Raman map of this mode account for the local variations of the indium content in the InGaN layers. Finally, the SO(A1) phonons of a single nanorod were detected, revealing a significantly reduced line width as compared to previous micro-Raman studies. The SO phonons of the InGaN layers were detected as a triplet structure which was only observed on the hexagonal side of the In-rich region. The carrier and surface sensitivity of tipenhanced Raman scattering was shown to provide detailed information on interfaces in nanostructures. The presented results demonstrate the full potential of this technique for a comprehensive analysis of next generation semiconductor nanostructures.

Table 1. Summary of Sample Properties Obtained by TERS Measurements and HRTEM Imaginga InGaN A1(TO) peak shift chemical composition clustering location polytypes planar defect stress carrier concentration

InN TO/ E1(TO) peak shift

GaN SO(A1) peak shift

InN TO/ E1(TO) peak intensity

HRTEM confirmation

X

X

X

X X

X X X X

X X X

The flags indicate the obtainable information based on the spatial analysis of peak shift and peak intensity of the different near field Raman modes and the confirmation by HRTEM. a

observed width of this mode in the TERS spectra is significantly smaller than usually reported in literature. This is caused by the fact that all previous works exclusively observe the SO modes in an ensemble of nanorods resulting in large inhomogeneous broadening. In contrast, the near-field signal of the TERS avoids the overlap of simultaneous contributions from different spatial regions, thus leading to narrower SO modes. Apparently, the GaN SO(A1) mode shifts by about 10 cm−1 between different spatial positions near the tip of a single nanorod as visualized by the Raman map in Figure 5c. The question remains as to the origin of the observed shift. The pronounced valley along the third and fourth row of the SO(A1) map reflects the locally reduced phonon energies of this surface-optical mode on the opposite side of the interface as compared to the strain-induced minimum of the A1(TO) InGaN mode (sixth row) in Figure 5f. This can be explained within the following model: caused by the strong electron− SO−phonon coupling in the thin GaN cap layer, the SO(A1) phonon frequency strongly depends on the carrier concentration.51,52 Considering that the interface between the two different polymorphs perpendicular to the cap layer creates a local perturbation in the charge density, the shift of the SO mode to lower phonon energies can be attributed to a local change in the carrier concentration. As calculated by Alencar et al. in ref 51, this shift is most likely related to a depletion zone induced by the underlying charge accumulation at the interface. Based on these arguments, the distribution of the local free carrier concentration is schematically visualized in the model in Figure 5d. In addition to the discussed GaN SO mode, up to three sharp lines near 700 cm−1 are observed at different positions. Taking the typical line width of reported ensemble spectra in GaN based materials into account,53 these lines would still lie within the frequency range of the broad band of GaN SO phonons. However, the observed spectral features in the TERS spectra are only detectable on the h-InN cluster side (see Supporting Information) and change their shape and frequency without clear pattern over the entire Raman map. Consequently, we do not attribute these line to the GaN capping layer but rather to SO phonon modes of the InGaN layers47,54,55 where the local indium fluctuations are responsible for the variation of the mode frequencies. In conclusion, a multitude of local properties with high spatial resolution is obtained based on tip-enhanced Raman

■

ASSOCIATED CONTENT

S Supporting Information *

InGaN SO phonon intensity, raw data of the line scans by TERS, TEM image of two dimensional vertical defect, and technical details and experimental method. This material is available free of charge via Internet at http://pubs.acs.org

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the European Research Council (ERC) under grant number 259286 and by the Deutsche Forschungsgemeinschaft (DFG) within the SFB 787. M.M. and M.S. gratefully acknowledge the support from the EC funded FP7 projects SMASH (no. 228999) and Gecco (280694). Support of the Spanish MINECO project TAPHOR (MAT2012-31392) and the AGAUR project SGR 150 is acknowledged.

■

REFERENCES

(1) Khan, A. Nat. Photonics 2009, 3, 432−434. (2) Qian, F.; Li, Y.; Gradecak, S.; Park, H.-G.; Dong, Y.; Ding, Y.; Wang, Z. L.; Lieber, C. M. Nat. Mater. 2008, 7, 701−6. (3) Guo, W.; Zhang, M.; Banerjee, A.; Bhattacharya, P. Nano Lett. 2010, 10, 3355−9. (4) Yeh, T.-W.; Lin, Y.-T.; Stewart, L. S.; Dapkus, P. D.; Sarkissian, R.; O’Brien, J. D.; Ahn, B.; Nutt, S. R. Nano Lett. 2012, 12, 3257−62. (5) Kim, H.-M.; Cho, Y.-H.; Lee, H.; Kim, S. I.; Ryu, S. R.; Kim, D. Y.; Kang, T. W.; Chung, K. S. Nano Lett. 2004, 4, 1059−1062. 3211


Nano Letters

Letter

(6) Monemar, B.; Sernelius, B. E. Appl. Phys. Lett. 2007, 91, 181103. (7) Colby, R.; Liang, Z.; Wildeson, I. H.; Ewoldt, D. A.; Sands, T. D.; García, R. E.; Stach, E. A. Nano Lett. 2010, 10, 1568−73. (8) Böhmler, M.; Wang, Z.; Myalitsin, A.; Mews, A.; Hartschuh, A. Angew. Chem., Int. Ed. 2011, 50, 11536−8. (9) Bao, W.; Melli, M.; Caselli, N.; Riboli, F.; Wiersma, D. S.; Staffaroni, M.; Choo, H.; Ogletree, D. F.; Aloni, S.; Bokor, J.; Cabrini, S.; Intonti, F.; Salmeron, M. B.; Yablonovitch, E.; Schuck, P. J.; WeberBargioni, A. Science 2012, 338, 1317−21. (10) Zhao, H.; Liu, G.; Zhang, J.; Poplawsky, J. D.; Dierolf, V.; Tansu, N. Opt. Express 2011, 19 (Suppl 4), A991−A1007. (11) Shen, Y. C.; Mueller, G. O.; Watanabe, S.; Gardner, N. F.; Munkholm, A.; Krames, M. R. Appl. Phys. Lett. 2007, 91, 141101− 141103. (12) Schubert, M. F.; Xu, J.; Kim, J. K.; Schubert, E. F.; Kim, M. H.; Yoon, S.; Lee, S. M.; Sone, C.; Sakong, T.; Park, Y. Appl. Phys. Lett. 2008, 93, 41102. (13) Ni, X.; Li, X.; Lee, J.; Liu, S.; Avrutin, V.; Ozgur, U.; Morkoc, H.; Matulionis, A.; Paskova, T.; Mulholland, G.; Evans, K. R. Appl. Phys. Lett. 2010, 97, 31110. (14) Nguyen, H. P. T.; Cui, K.; Zhang, S.; Djavid, M.; Korinek, A.; Botton, G. A.; Mi, Z. Nano Lett. 2012, 12, 1317−23. (15) Cho, H. K.; Lee, J. Y.; Sharma, N.; Humphreys, C. J.; Yang, G. M.; Kim, C. S.; Song, J. H.; Yu, P. W. Appl. Phys. Lett. 2001, 79, 2594. (16) Jinschek, J. R.; Erni, R.; Gardner, N. F.; Kim, A. Y.; Kisielowski, C. Solid State Commun. 2006, 137, 230−234. (17) Yeo, B.-S.; Stadler, J.; Schmid, T.; Zenobi, R.; Zhang, W. Chem. Phys. Lett. 2009, 472, 1−13. (18) Berweger, S.; Neacsu, C. C.; Mao, Y.; Zhou, H.; Wong, S. S.; Raschke, M. B. Nat. Nanotechnol. 2009, 4, 496−9. (19) Irmer, G.; Monecke, J.; Verma, P. Light Scattering of Semiconducting Nanoparticles. Encyclopedia of Nanoscience and Nanotechnology; American Scientific Publishers: Valencia, CA, 2004; Vol. X, pp 561−586. (20) Matsui, R.; Verma, P.; Ichimura, T.; Inouye, Y.; Kawata, S. Appl. Phys. Lett. 2007, 90, 061906. (21) Jiang, N.; Foley, E. T.; Klingsporn, J. M.; Sonntag, M. D.; Valley, N. A.; Dieringer, J. A.; Seideman, T.; Schatz, G. C.; Hersam, M. C.; Van Duyne, R. P. Nano Lett. 2012, 12, 5061−7. (22) Stadler, J.; Schmid, T.; Zenobi, R. Nano Lett. 2010, 10, 4514− 20. (23) Stockle, R. M.; Suh, Y. D.; Deckert, V.; Zenobi, R. Chem. Phys. Lett. 2000, 318, 131−136. (24) Weber-Bargioni, A.; Schwartzberg, A.; Cornaglia, M.; Ismach, A.; Urban, J. J.; Pang, Y.; Reuven, G.; Bokor, J.; Salmeron, M. B.; Ogletree, D. F.; Ashby, P.; Cabrini, S.; Schuck, P. J. Nano Lett. 2011, 11, 1201−1207. (25) Schmid, T.; Opilik, L.; Blum, C.; Zenobi, R. Angew. Chem., Int. Ed. 2013, 52, 2−17. (26) Sevinc, P. C.; Wang, X.; Wang, Y.; Zhang, D.; Meixner, A. J.; Lu, H. P. Nano Lett. 2011, 11, 1490−4. (27) Wang, X.; Li, S.; Fündling, S.; Wei, J.; Erenburg, M.; Ledig, J.; Wehmann, H. H.; Waag, A.; Bergbauer, W.; Mandl, M.; Strassburg, M.; Steegmüller, U. Int. J. High Speed Electron. Syst. 2012, 21, 1250008. (28) Eligal, L.; Culfaz, F.; McCaughan, V.; Cade, N. I.; Richards, D. Rev. Sci. Instrum. 2009, 80, 033701. (29) Ren, B.; Picardi, G.; Pettinger, B. Rev. Sci. Instrum. 2004, 75, 837. (30) Hayashi, S. Top. Appl. Phys. 2001, 81, 71−95. (31) Okamoto, K.; Niki, I.; Shvartser, A.; Narukawa, Y.; Mukai, T.; Scherer, A. Nat. Mater. 2004, 3, 601−5. (32) Harima, H. J. Phys.: Condens. Matter 2002, 14, R967−R993. (33) Siegle, H.; Kaczmarczyk, G.; Filippidis, L.; Litvinchuk, A. P.; Hoffmann, A.; Thomsen, C. Phys. Rev. B 1997, 55, 7000−7004. (34) Callsen, G.; Reparaz, J. S.; Wagner, M. R.; Kirste, R.; Nenstiel, C.; Hoffmann, A.; Phillips, M. R. Appl. Phys. Lett. 2011, 98, 061906. (35) Kaczmarczyk, G.; Kaschner, A.; Reich, S.; Hoffmann, A.; Thomsen, C.; As, D. J.; Lima, A. P.; Schikora, D.; Lischka, K.; Averbeck, R.; Riechert, H. Appl. Phys. Lett. 2000, 76, 2122.

(36) Tabata, A.; Lima, A. P.; Teles, L. K.; Scolfaro, L. M. R.; Leite, J. R.; Lemos, V.; Schöttker, B.; Frey, T.; Schikora, D.; Lischka, K. Appl. Phys. Lett. 1999, 74, 362. (37) Cho, H. K.; Lee, J. Y.; Kim, C. S.; Yang, G. M.; Sharma, N.; Humphreys, C. J. Cryst. Growth 2001, 231, 466−473. (38) Duxbury, N.; Bangert, U.; Dawson, P.; Thrush, E. J.; Van der Stricht, W.; Jacobs, K.; Moerman, I. Appl. Phys. Lett. 2000, 76, 1600. (39) Schörmann, J.; As, D. J.; Lischka, K.; Schley, P.; Goldhahn, R.; Li, S. F.; Löffler, W.; Hetterich, M.; Kalt, H. Appl. Phys. Lett. 2006, 89, 261903. (40) Schley, P.; Goldhahn, R.; Winzer, A.; Gobsch, G.; Cimalla, V.; Ambacher, O.; Lu, H.; Schaff, W.; Kurouchi, M.; Nanishi, Y.; Rakel, M.; Cobet, C.; Esser, N. Phys. Rev. B 2007, 75, 205204. (41) King, P. D. C.; Veal, T. D.; McConville, C. F.; Fuchs, F.; Furthmüller, J.; Bechstedt, F.; Schley, P.; Goldhahn, R.; Schörmann, J.; As, D. J.; Lischka, K.; Muto, D.; Naoi, H.; Nanishi, Y.; Lu, H.; Schaff, W. J. Appl. Phys. Lett. 2007, 91, 092101. (42) Reparaz, J. S.; Peica, N.; Kirste, R.; Goni, A.; Wagner, M. R.; Callsen, G.; Alonso, M. I.; Garriga, M.; Carmen Marcus, I.; Ronda, A.; Berbezier, I.; Maultzsch, J.; Thomsen, C.; Hoffmann, A. Nanotechnology 2013, 24, 185704. (43) Grille, H.; Schnittler, C.; Bechstedt, F. Phys. Rev. B 2000, 61, 6091−6105. (44) Tabata, A.; Leite, J. R.; Lima, A. P.; Silveira, E.; Lemos, V.; Frey, T.; As, D. J.; Schikora, D.; Lischka, K. Appl. Phys. Lett. 1999, 75, 1095. (45) Kontos, A. G.; Raptis, Y.; Pelekanos, N.; Georgakilas, A.; BelletAmalric, E.; Jalabert, D. Phys. Rev. B 2005, 72, 155336. (46) Goñi, A.; Siegle, H.; Syassen, K.; Thomsen, C.; Wagner, J.-M. Phys. Rev. B 2001, 64, 035205. (47) Zhu, J. H.; Ning, J. Q.; Zheng, C. C.; Xu, S. J.; Zhang, S. M.; Yang, H. Appl. Phys. Lett. 2011, 99, 113115. (48) Dhara, S.; Sahoo, P.; Tyagi, A.; Raj, B. Surface Optical Mode in Semiconductor Nanowires. In Nanowires - Implementation and Applications, Vol. 2; Intech Open Access: Croatia, 2011. (49) Mata, R.; Cros, A.; Hestroffer, K.; Daudin, B. Phys. Rev. B 2012, 85, 035322. (50) Wang, J.; Demangeot, F.; Péchou, R.; Ponchet, A.; Cros, A.; Daudin, B. Phys. Rev. B 2012, 85, 155432. (51) Alencar, A. M.; Filho, R. N. C.; Freire, V. N.; Farias, G. A. Phys. Rev. B 1998, 57, 12275. (52) Alencar, A.; Freire, V.; Farias, G. Braz. J. Phys. 1997, 27/A, 240. (53) Wu, S.-E.; Dhara, S.; Hsueh, T.-H.; Lai, Y.-F.; Wang, C.-Y.; Liu, C.-P. J. Raman Spectrosc. 2009, 40, 2044−2049. (54) Dutta, M.; Alexson, D.; Bergman, L.; Nemanich, R. J.; Dupuis, R.; Kim, K. W.; Komirenko, S.; Stroscio, M. Physica E: LowDimensional Syst.Nanostruct. 2001, 11, 277−280. (55) Wang, X. H.; Ning, J. Q.; Xu, S. J.; Choi, H. W. J. Appl. Phys. 2011, 110, 093111.

3212


Nanoscale Imaging of InN Segregation and ... - ACS Publications

Recommend Documents