GaAs Nanowires for Near-Infrared

Mar 28, 2014 - Adam C. Scofield , Andrew Lin , Michael Haddad , and Diana L. Huffaker .... Michael Hanke , Steven John Leake , Otmar Loffeld , Ullrich...
1 downloads 0 Views 4MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Coaxial Multishell (In,Ga)As/GaAs Nanowires for Near-Infrared Emission on Si Substrates Emmanouil Dimakis,* Uwe Jahn, Manfred Ramsteiner, Abbes Tahraoui, Javier Grandal, Xiang Kong, Oliver Marquardt, Achim Trampert, Henning Riechert, and Lutz Geelhaar Paul-Drude-Institut für Festkörperelektronik, Hausvogteiplatz 5−7, 10117 Berlin, Germany ABSTRACT: Efficient infrared light emitters integrated on the mature Si technology platform could lead to on-chip optical interconnects as deemed necessary for future generations of ultrafast processors as well as to nanoanalytical functionality. Toward this goal, we demonstrate the use of GaAs-based nanowires as building blocks for the emission of light with micrometer wavelength that are monolithically integrated on Si substrates. Freestanding (In,Ga)As/GaAs coaxial multishell nanowires were grown catalyst-free on Si(111) by molecular beam epitaxy. The emission properties of single radial quantum wells were studied by cathodoluminescence spectroscopy and correlated with the growth kinetics. Controlling the surface diffusivity of In adatoms along the NW side-walls, we improved the spatial homogeneity of the chemical composition along the nanowire axis and thus obtained a narrow emission spectrum. Finally, we fabricated a light-emitting diode consisting of approximately 105 nanowires contacted in parallel through the Si substrate. Room-temperature electroluminescence at 985 nm was demonstrated, proving the great potential of this technology. KEYWORDS: semiconductor, core-shell, light emitting diode, molecular beam epitaxy

T

spectral range can be accessed by employing narrow band gap III-V semiconductors as the active medium. One sensible choice would be the (In,Ga)As ternary alloys, which have been widely used in thin film LEDs, because their band gaps span from the near-IR to the mid-IR range depending on the alloy composition. However, (In,Ga)As NWs suffer from a high number density of planar defects when grown in the vapor− solid mode,12−16 while they exhibit spatially inhomogeneous composition17−19 or limited incorporation of In20 when grown in the vapor−liquid−solid (VLS) mode, making the fabrication of LEDs with axial quantum heterostructures challenging. Only a remotely similar approach using core/multishell (In,Ga)As/ GaAs nanoneedles has been employed for the realization of LEDs with emission at 1.128 μm.21 The nanoneedles were grown on a previously roughened Si substrate, while the quantum well (QW) was formed on the tapered sidewalls. In this Letter, we employ a coaxial multishell scheme for the realization of NWs that contain a single radial (In,Ga)As/GaAs QW. These NWs are grown on Si by molecular beam epitaxy (MBE) without the use of any foreign catalyst particles in order to avoid the contamination of the underlying Si platform and the NWs themselves. We investigate the emission properties of the NWs in conjunction with the growth kinetics, retrieving in that way information about the structural homogeneity of the radial QWs and how it is affected by the growth conditions. This information is used for the optimization of the emission features of the radial QWs before their incorporation in the

he integration of logic and light on the same chip would advance the current complementary-metal-oxide-semiconductor (CMOS) technology based on Si by enhancing performance and adding new functionalities. Possible applications include, for example, processor chips benefitting from faster data transfer through optical interconnects as well as autonomous systems powered by photovoltaics and capable of optical nanoanalytics.1−4 The generation of light in Si, though, has been hindered by the indirect nature of the energy band gap of Si. A very promising approach is based on the fact that III-V semiconductors with direct band gap can be grown epitaxially in high crystal quality on Si in the form of nanowires (NWs), despite the large mismatch in lattice and thermal expansion coefficients between the two material systems. Thus, III-V light emitters could be monolithically integrated on the underlying Si CMOS platform and could be exploited for the conversion of electrical signals into optical ones. The inverse process could also be realized with III-V NWs as light detectors, offering, in fact, a wider spectral response than Si photonics. The simplest device for the conversion of electrical signals into optical ones is a light emitting diode (LED). To date, III-V LEDs on Si have been demonstrated using (In,Ga)N/GaN,5−7 GaN/(Al,Ga)N,8 GaAs/(In,Ga)P,9 GaAs/(Al,Ga)As,10 and Ga(As,P)/GaAs11 quantum heterostructure NWs. Ultraviolet, blue, green, and red emission has been obtained with the GaNbased NWs, while those based on GaAs provide access to the near-infrared (IR) range, and the longest emission wavelength reported is about 830 nm. Arguably the most relevant application could be optical data transfer, and the underlying optical fiber technology requires LED operation at the telecommunication wavelengths 1.3 and 1.5 μm. This particular © 2014 American Chemical Society

Received: February 3, 2014 Revised: March 14, 2014 Published: March 28, 2014 2604

dx.doi.org/10.1021/nl500428v | Nano Lett. 2014, 14, 2604−2609

Nano Letters

Letter

Figure 1. (a) Side-view SEM image of (In,Ga)As/GaAs SQW-NWs grown on Si(111). The arrow indicates the parasitic layer. (b) Plan-view TEM image of a single NW showing the coaxial heterostructure. The core and the various shells are indicated. (c) SAD pattern acquired along the [111] zone axis. The scale bar corresponds to 3 μm in (a) and 50 nm in (b).

grown on Si(111) with a number density sufficiently low (5−10 × 107 cm−2) so that the beam shadowing between neighboring NWs during growth was limited (all molecular/atomic beams impinge onto the substrate in an angle of 30° with respect to the substrate normal). Because of the kinetically limited growth conditions of the NW shell layers, unintentional overgrowth of the substrate also took place as indicated with the arrow. Specifically, for every NW shell layer of a given thickness t growing intentionally on the NW side-walls, a [t(π/tan30°) = 5.4t]-thick planar layer (henceforth referred to as parasitic layer) is also growing directly on the highly mismatched Si substrate. However, these parasitic layers (considering also the presence of the Si-oxide layer at the interface with the substrate) are highly defective and thus optically inactive as will be demonstrated in the LED section. The microstructure of the radial QW was investigated by plan-view transmission electron microscopy (TEM). The preparation of the TEM specimens consisted of the mechanical stabilization of the as-grown NWs on their original substrate using an epoxy adhesive resin, and the subsequent thinning of the specimens from the side of the substrate using standard methods of grinding, dimpling, and Ar ion milling until electron transparency was obtained. This methodology allowed the investigation of multiple NWs within each TEM specimen. As shown in Figure 1b, the multishell structure around the 50 nm thick GaAs core consists of a 50 nm thick inner GaAs shell (this is indicated artificially, because it cannot be distinguished from the GaAs core), an 11 nm thick (In,Ga)As intermediate shell that appears as a well-defined hexagonal stripe with dark image contrast, and finally, a 30 nm thick outer GaAs shell. The role of the inner GaAs shell was to increase the diameter of the (In,Ga)As shell for comparability with the LED structure that will be presented in the following. The NWs have the ZB structure with {11̅0}-faceted sidewalls as determined by complementary TEM imaging and selected area diffraction

active region of LED structures. Finally, we demonstrate roomtemperature operation at 985 nm for an LED consisting of an ensemble of NWs contacted in parallel through the Si substrate. These results show the great potential of the hybrid technology comprising III-V NWs on Si for the operation at even longer wavelengths. Results and Discussion. All the coaxial multishell (In,Ga)As/GaAs NWs investigated here were grown catalystfree on p+ −Si(111) substrates covered with their native oxide by solid-source MBE following a two-stage procedure. First, undoped GaAs NWs were grown on Si to serve as cores for the subsequent conformal overgrowth with multiple coaxial shells of different composition and/or doping. The GaAs cores were grown at 580 °C with a V/III flux ratio of 13, resulting in 3.5 μm long NWs after 15 min of growth. Their growth proceeds in the VLS mode with a liquid Ga droplet driving the axial growth along ⟨111⟩. At the end of the core-growth stage, the Ga droplets were converted to GaAs by exposure to As4 in order to prevent the further elongation of the NWs during the subsequent shell-growth stage. Then, the NW shells were grown conformally around the cores at 390 °C with a V/III flux ratio of 82. Under these conditions, we limit the surface diffusivity of the Ga adatoms on the {11̅0} NW side-walls (sticking coefficient equal to 1) thus achieving a homogeneous shell thickness along the NW axis.22 The aforementioned growth temperatures refer to temperatures of the Si substrate as measured by optical pyrometry. NWs with a single radial QW (SQW-NWs) were obtained by sandwiching one (In,Ga)As shell layer between two GaAs ones. The LED structures involved in addition the growth of GaAs shells doped with either Si donors or Be acceptors on either side of the single radial QW. The morphology of an as-grown sample with SQW-NWs is illustrated in the side-view scanning electron microscopy (SEM) image of Figure 1a. Free-standing SQW-NWs were 2605

dx.doi.org/10.1021/nl500428v | Nano Lett. 2014, 14, 2604−2609

Nano Letters

Letter

(SAD) investigations. In the SAD pattern taken along the [111] zone axis and shown in Figure 1c, the diffraction spots from the different shells cannot be clearly resolved, suggesting that the (In,Ga)As QW has a high degree of strain with respect to the inner GaAs shell. In fact, because no misfit dislocations in highresolution TEM were observed, we can assume that the (In,Ga) As shell was grown pseudomorphically. The In-content (xIn) of the (In,Ga)As shell was measured by electron energy-loss spectroscopy (EELS) in the TEM using a plan-view specimen. However, due to the resolution of our system, it was measured in SQW-NWs with a thicker QW and assumed to be similar to the case of a QW thickness of 11 nm, because the growth conditions were nominally the same. xIn was found to be in a range of 22 ± 3 to 28 ± 3%, which is in good agreement with the Ga/In flux ratio of 3:1 that we used. These composition variations as well as thickness variations up to 3 nm were found between the radial QWs in different NWs or on different side-walls of the same NW and may be attributed to unintentional deviations in the growth conditions along the substrate surface and/or the beam shadowing (although limited) from the surrounding NWs. A more detailed high-resolution TEM investigation is currently in progress, and the results will be published elsewhere. The emission properties of the (In,Ga)As/GaAs SQW-NWs were investigated by cathodoluminescence (CL) spectroscopy, using a series of four SQW-NW samples with different QW thickness (LQW = 4, 7, 11, and 13 nm). The measurements were performed on collections of NWs at 7 K in a field-emission SEM using a photomultiplier optimized for the near-IR range (InP/(In,Ga)As photocathode with spectral range of 300−1700 nm). The NWs had been previously harvested from their substrates and dispersed onto Si carrier wafers (which were covered with a thin layer of Au to avoid charging effects), forming mixed collections of discrete as well as bundled NWs. In that way, we can be sure that the CL measurements are not affected by the presence of the parasitic layers. As shown in Figure 2a, all CL spectra consist of a single peak that redshifts with increasing LQW as expected for radiative recombination of electron−hole pairs inside a QW. The asymmetry that characterizes the CL line shape for the thinnest QW disappears gradually with increasing LQW. We speculate that this asymmetry originates from small xIn and LQW variations within the same and/or between different NWs. Their effect on the emission features would be weaker for thicker (In,Ga)As shells, which is in agreement with the experiment. We have simulated the dependence of the ground state recombination in (In,Ga)As/GaAs QWs on LQW using a onedimensional plane-wave based eight-band k·p model. The excitonic emission has been neglected due to the small exciton binding energy. The model of a QW infinitely extended in one plane appears reasonable here, as the diameter of the (In,Ga)As shell is much larger than its thickness, such that contributions to energy quantization are a result of the shell thickness rather than of its diameter. The elastic properties were determined using a linear elasticity model and were taken into account in the k·p model.23,24 The material parameters employed were taken from ref 25. As shown in Figure 2b, remarkable agreement between the CL recombination energies (symbols) and the theoretically predicted ones (curve) was obtained assuming a coherently strained (In,Ga)As shell with xIn = 25%. This not only proves that the emission originates from the recombination of electron−hole pairs inside the radial QW but also demonstrates the good control that we have on the growth

Figure 2. (a) CL spectra from (In,Ga)As/GaAs SQW-NW collections with different LQW at 7 K, all grown at Tshell = 390 °C. (b) Calculated ground-state recombination wavelength in an (In,Ga)As/GaAs QW as a function of LQW (black curve), and comparison with the CL peak position (symbols) measured in (a). One coherently strained QW with xIn = 25% was assumed in the calculation. The shaded area represents a variation of xIn in the calculation up to ±1%.

of coaxial multishell (In,Ga)As/GaAs NWs. To emphasize the level of agreement between experiment and theory, the theoretical curve includes a spread of ±1% for xIn, shown as the shadowed area. The spatial homogeneity of the light emission along the axis of a single SQW-NW is demonstrated by the CL spectral line scan in Figure 3a (a Si-CCD detector was used for this CL mode). The emission from the QW (11 nm thick in this case) is centered at 955 nm for the most part of the NW length. Only a small blueshift by 5 nm is observed at the bottom 1 μm of the NW, which corresponds to less than 1% difference in xIn. Finally, a CL signal is never emitted from the NW tip at any wavelength. This happens because the crystal quality of all shell layers in this region is dictated by the highly defective tip of the GaAs core, which was formed during the exposure of the Ga droplet to As4 at the end of the core-growth stage as observed by in situ reflection high energy electron diffraction. Similar observations have been made by other groups using TEM.26 The spectral width of the emitted light from the main part of the NW (fwhm = 20 nm/28 meV) is most likely dictated by the small xIn and LQW variations between the different NW sidewalls that we observed by TEM. As expected, when a collection of the same type of NWs is measured by CL (see the spectrum in Figure 2a for LQW = 11 nm), xIn and LQW variations between different NWs cause a larger broadening of the line shape (fwhm = 50 nm/67 meV). The sufficiently low Tshell (390 °C) is the key parameter in order to obtain spatially homogeneous light emission, which in turn means homogeneous xIn and LQW of the (In,Ga)As shell 2606

dx.doi.org/10.1021/nl500428v | Nano Lett. 2014, 14, 2604−2609

Nano Letters

Letter

position at 962 nm for Tshell = 390 °C is consistent with the nominal xIn = 25%. The emission peak at 938 nm for Tshell = 415 °C and at 910 nm for Tshell = 430 °C corresponds to xIn of approximately 22 and 18%, respectively. Noticeably, the longwavelength emission tails extend to wavelengths that correspond to xIn > 25%. Considering the kinetic limitations of the growth at these low Tshell, we believe that all the observed variations of xIn originate from the increase of the surface diffusivity of the In adatoms along the NW side-walls with increasing Tshell. In that way, In adatoms are transferred from the NW side-walls to the parasitic layer or concentrate close to the NW tips. Thus, it is essential to limit the surface diffusivity of In adatoms in order to obtain spatially homogeneous xIn. Having identified the growth conditions that lead to SQWNWs with improved emission properties, we grew a complete LED structure. The cross-section of the LED NWs is sketched in Figure 4a, where the core and each shell are shown with

Figure 3. (a) CL spectral line scan along the axis of a single (In,Ga)As/GaAs SQW-NW with 11 nm thick QW, grown at Tshell = 390 °C. The linear color scale is given in arbitrary units. Inset: SEM image of the measured NW in-scale with the CL plot. (b) CL spectra from (In,Ga)As/GaAs SQW-NW collections grown at three different Tshell (390, 415, and 430 °C). Inset: False-color monochromatic CL maps for two (In,Ga)As/GaAs SQW-NWs grown at Tshell = 415 and 430 °C, showing the spatial distribution of the CL intensity at the sample wavelengths indicated by the arrows. The NW length was 3.5 μm in all cases. All CL measurements in (a,b) were performed at 7 K.

along the NW axis. The effect of the growth temperature on the emission features is clearly demonstrated in Figure 3b, where the CL spectra from SQW-NW collections (LQW = 11 nm) grown at Tshell = 390, 415, and 430 °C are plotted. Even though the SQW-NWs did not show any morphological differences with respect to each other, their emission peak blue-shifted with increasing Tshell, while broad emission features developed on its long-wavelength side. Spatially resolved monochromatic CL measurements on single NWs showed that the main CL peak at any Tshell is emitted homogeneously from the most part of the NW axis, as in the case of Tshell = 390 °C discussed in Figure 3a. However, the broad long-wavelength emission that was observed for Tshell = 415 and 430 °C originates only from the vicinity of the NW tips. This can be seen in the false-color monochromatic CL maps in Figure 3b, where the emission at given sample wavelengths (indicated by arrows) is probed. At this low Tshell range, where the surface diffusion of the Ga adatoms and the Ga/In intermixing are strongly suppressed, the CL emission at different wavelengths must originate from QWs with different xIn. That is, the blueshift of the main CL peak is caused by a decrease of xIn with increasing Tshell, while the broad emission at longer wavelengths is attributed to increased xIn close to the NW tips. In fact, calculations of the ground state recombination in these QWs showed that only the CL peak

Figure 4. (a) Cross-sectional schematic of the (In,Ga)As/GaAs LED NWs. (b) False-color monochromatic CL map of the as-grown LED sample at 7 K. The probe wavelength was 940 nm. (c) Side-view SEM image of the fully processed LED. Inset: Top-view optical image of the LED. The scale bar corresponds to 1 μm in (b,c) and 100 μm in the inset of (c).

different colors depending on their composition and doping. Starting from the core and moving toward the outer shell, the LED structure consists of (1) an undoped GaAs core, (2) a 55 nm thick p-type GaAs:Be shell, (3) one 9 nm thick undoped GaAs shell on each side of the QW, 4) an 11 nm thick (In,Ga) As QW shell with nominal xIn = 25%, and 5) a 55 nm thick ntype GaAs:Si shell. Details about the shell doping of GaAs NWs with Si for controlled n-type conductivity can be found in ref 22. 2607

dx.doi.org/10.1021/nl500428v | Nano Lett. 2014, 14, 2604−2609

Nano Letters

Letter

The emission properties were first assessed by spatially resolved monochromatic CL. The false-color CL map in Figure 4b shows that light was emitted only from the radial QWs inside the NWs, while no emission was recorded from the highly defective parasitic layers (where the carrier recombination via defect-states is nonradiative). As probe wavelength, we chose the peak wavelength (940 nm) of the full CL spectrum (not shown here). The emission peak is blue shifted with respect to the 962 nm of the SQW-NWs with the same xIn and LQW. We believe this is due to the higher absorption of heat in the p-type GaAs shell compared to the undoped one during growth, which results in higher Tshell for the LED compared to the SQW-NWs. Figure 4c shows a side-view SEM image of the fully processed device. The n-type shell was contacted from the front side of the device with sputtered indium tin oxide (ITO), while the p-type shell was contacted through the p+−Si substrate with an aluminum back-contact. A 1.5 μm thick benzocyclobutene (BCB) layer was employed to electrically insulate the ITO contact from the substrate.27 The transparency of both the BCB and the ITO ensured the unobstructed propagation/ escape of the emitted light. Furthermore, all epitaxial structures with lower height than the NWs remained fully embedded into the BCB layer without influencing the device operation. Finally, a cross-hair contact made of Au was deposited on top of the ITO in order to improve the current distribution over the whole device area. The top-view of the LED tested in the following is shown as an inset in Figure 4c. The current−voltage (I−V) characteristic presented rectifying behavior as illustrated in Figure 5a. The leakage current is in the range of a few μA for a reverse bias of −10 V, which is satisfactory considering the ambiguous quality of the p−n junction at the defective NW tips, where any CL signal was never detected, and the NW bases, where the shells are in contact with the defective parasitic layers. The turn-on voltage of 5 V is relatively high, but a considerable drop of voltage must occur at the GaAs/Si interface, where the existence of a thin Sioxide layer between GaAs and Si has been reported in TEM investigations.28 The Si-oxide layer should also be the source of the series resistance that limits the current at high voltages. All these issues may be alleviated by using prepatterned substrates, where the topography of the Si-oxide layer on Si can be predefined, and oxide-free GaAs/Si interfaces may be obtained. The room-temperature electroluminescence (EL) spectrum of the LED is shown in Figure 5b. The device was driven by +15 V voltage pulses at 933 kHz with 20% duty cycle, while the emission was recorded with a Si-CCD detector with cutoff at 1100 nm. The peak position of the emission at 985 nm is consistent with the nominal xIn and LQW of the (In,Ga)As shell at room temperature, indicating that the injected carriers recombine inside the radial QW. The false-color EL map in the inset shows that the light is emitted from localized centers, as would be expected for emission from the NWs, without being a direct evidence though that the emission does not come from the extended parasitic layers. Anyway, we can be sure that this cannot be the case because the parasitic (In,Ga)As layer is 60 nm thick, and thus fully relaxed, and should emit at much longer wavelengths. Variations in the emission properties and the biasing conditions of every single NW of the same device would increase the EL spectral width. In fact, the large spectral width of the LED measured here (fwhm = 100 nm/128 meV) can be considered reasonable, because the device consists of approximately 105 parallel contacted NWs (assuming that all of

Figure 5. (a) Current−voltage characteristic of the (In,Ga)As/GaAs NW LED. Inset: The same characteristic in linear scale. (b) The corresponding EL spectrum at room-temperature. Inset: Lowresolution EL map (false-color) showing the spatial distribution of the EL intensity over the LED area.

them have been contacted successfully). It is anticipated that growth of such NW LEDs on prepatterned substrates would minimize the variations between different NWs and thus the width of the EL spectrum. Aiming at telecommunication wavelengths, QWs with a higher xIn would be necessary. In terms of planar heterostrustures, this would cause the formation of misfit dislocations due to the large misfit strain between the (In,Ga)As and the GaAs layers. However, the coaxial multishell NWs appear as a possible solution due to the unique accommodation of the misfit strain in the core−shell geometry.29−31 Conclusions. In conclusion, coaxial multishell (In,Ga)As/ GaAs NWs were grown on Si and studied as light emitters. Their emission properties were correlated with the growth kinetics, showing that proper control of the surface diffusivity of the group-III adatoms leads to spatially homogeneous heterostructures with the desired emission characteristics. A LED with room-temperature emission at 985 nm was fabricated by contacting an ensemble of free-standing NWs on Si. Our results demonstrate the great potential of (In,Ga)As/GaAs NWs for near-IR light emission on a Si platform.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] Present Address

(E.D.) Helmholtz-Zentrum Dresden-Rossendorf, Bautzner Landstraße 400, 01328 Dresden, Germany. Notes

The authors declare no competing financial interest. 2608

dx.doi.org/10.1021/nl500428v | Nano Lett. 2014, 14, 2604−2609

Nano Letters

■

Letter

(31) Nazarenko, M. V.; Sibirev, N. V.; Wei, N. K.; Ren, F.; Son, K. W.; Dubrovskii, V. G.; Chang-Hasnain, C. J. Appl. Phys. 2013, 113, 104311.

ACKNOWLEDGMENTS The authors thank Jacob Dinner, Bernd Drescher, and Walid Anders for the device processing, Klaus Biermann for preliminary k·p calculations, and Claudia Herrmann for the MBE maintenance.

■

REFERENCES

(1) Yang, P.; Yan, R.; Fardy, M. Nano Lett. 2010, 10, 1529−36. (2) Lieber, C. M. MRS Bull. 2011, 36, 1052−1063. (3) Borgström, M. T.; Wallentin, J.; Heurlin, M.; Fält, S.; Wickert, P.; Leene, J.; Magnusson, M. H.; Deppert, K.; Samuelson, L. IEEE J. Sel. Top. Quant. 2011, 17, 1050−1061. (4) Roelkens, G.; Liu, L.; Liang, D.; Jones, R.; Fang, A.; Koch, B.; Bowers, J. Laser Photonics Rev. 2010, 4, 751−779. (5) Kikuchi, A.; Kawai, M.; Tada, M.; Kishino, K. Jpn. J. Appl. Phys. 2004, 43, L1524−L1526. (6) Li, S.; Waag, A. J. Appl. Phys. 2012, 111, 071101. (7) Limbach, F.; Hauswald, C.; Lähnemann, J.; Wölz, M.; Brandt, O.; Trampert, A.; Hanke, M.; Jahn, U.; Calarco, R.; Geelhaar, L.; Riechert, H. Nanotechnology 2012, 23, 465301. (8) Sekiguchi, H.; Kishino, K.; Kikuchi, A. Electron. Lett. 2008, 44, 151−152. (9) Svensson, C. P. T.; Mårtensson, T.; Trägårdh, J.; Larsson, C.; Rask, M.; Hessman, D.; Samuelson, L.; Ohlsson, J. Nanotechnology 2008, 19, 305201. (10) Tomioka, K.; Motohisa, J.; Hara, S.; Hiruma, K.; Fukui, T. Nano Lett. 2010, 10, 1639−1644. (11) Tomioka, K.; Tanaka, T.; Hara, S.; Hiruma, K.; Fukui, T. IEEE J. Sel. Top. Quantum Electron. 2011, 17, 1112−1129. (12) Shin, J. C.; Kim, K. H.; Yu, K. J.; Hu, H.; Yin, L.; Ning, C.-Z.; Rogers, J. A.; Zuo, J.-M.; Li, X. Nano Lett. 2011, 11, 4831−4838. (13) Yoshimura, M.; Tomioka, K.; Hiruma, K.; Hara, S.; Motohisa, J.; Fukui, T. J. Cryst. Growth 2011, 315, 148−151. (14) Morkötter, S.; et al. Phys. Rev. B 2013, 87, 205303. (15) Shapiro, J. N.; Lin, A.; Wong, P. S.; Scofield, A. C.; Tu, C.; Senanayake, P. N.; Mariani, G.; Liang, B. L.; Huffaker, D. L. Appl. Phys. Lett. 2010, 97, 243102. (16) Tomioka, K.; Yoshimura, M.; Fukui, T. Nature 2012, 488, 189− 92. (17) Kim, Y.; Joyce, H. J.; Gao, Q.; Tan, H. H.; Jagadish, C.; Paladugu, M.; Zou, J.; Suvorova, A. A. Nano Lett. 2006, 6, 599−604. (18) Ye, H.; Yu, Z. Y.; Kodambaka, S.; Shenoy, V. B. Appl. Phys. Lett. 2012, 100, 263103. (19) Guo, Y.-N.; Xu, H.-Y.; Auchterlonie, G. J.; Burgess, T.; Joyce, H. J.; Gao, Q.; Tan, H. H.; Jagadish, C.; Shu, H.-B.; Chen, X.-S.; Lu, W.; Kim, Y.; Zou, J. Nano Lett. 2013, 13, 643−50. (20) Heiss, M.; Ketterer, B.; Uccelli, E.; Morante, J. R.; Arbiol, J.; Fontcuberta i Morral, A. Nanotechnology 2011, 22, 195601. (21) Chuang, L. C.; Sedgwick, F. G.; Chen, R.; Ko, W. S.; Moewe, M.; Ng, K. W.; Tran, T.-T. D.; Chang-Hasnain, C. Nano Lett. 2011, 11, 385−90. (22) Dimakis, E.; Ramsteiner, M.; Tahraoui, A.; Riechert, H.; Geelhaar, L. Nano Res. 2012, 5, 796−804. (23) Marquardt, O.; Schulz, S.; Freysoldt, C.; Boeck, S.; Hickel, T.; O’Reilly, E. P.; Neugebauer, J. Opt. Quantum. Electron. 2011, 44, 183− 188. (24) Marquardt, O.; Boeck, S.; Freysoldt, C.; Hickel, T.; Neugebauer, J. Comput. Phys. Commun. 2010, 181, 765−771. (25) Schliwa, A.; Winkelnkemper, M.; Bimberg, D. Phys. Rev. B 2007, 76, 205324. (26) Heon Kim, Y.; Woo Park, D.; Jun Lee, S. Appl. Phys. Lett. 2012, 100, 033117. (27) Chia, A. C. E.; LaPierre, R. R. Nanotechnology 2011, 22, 245304. (28) Biermanns, A.; Breuer, S.; Trampert, A.; Davydok, A.; Geelhaar, L.; Pietsch, U. Nanotechnology 2012, 23, 305703. (29) Kavanagh, K. L. Semicond. Sci. Technol. 2010, 25, 024006. (30) Haapamaki, C. M.; Baugh, J.; LaPierre, R. R. J. Appl. Phys. 2012, 112, 124305. 2609

dx.doi.org/10.1021/nl500428v | Nano Lett. 2014, 14, 2604−2609