Control over the Number Density and Diameter of GaAs Nanowires on

Jul 30, 2013 - Influence of growth parameters on In-droplet-assisted growth of InAs nanowires on silicon. Ezekiel A. Anyebe. Applied Nanoscience 2017 ...
0 downloads 0 Views 1MB Size
Letter pubs.acs.org/NanoLett

Control over the Number Density and Diameter of GaAs Nanowires on Si(111) Mediated by Droplet Epitaxy Claudio Somaschini,*,† Sergio Bietti,‡ Achim Trampert,† Uwe Jahn,† Christian Hauswald,† Henning Riechert,† Stefano Sanguinetti,‡ and Lutz Geelhaar† †

Paul-Drude-Institut für Festkörperelektronik, Hausvogteiplatz 5−7, 10117 Berlin, Germany LNESS and Dipartimento di Scienza dei Materiali dell’Università di Milano-Bicocca, via Cozzi 53, 20125 Milano, Italy



S Supporting Information *

ABSTRACT: We present a novel approach for the growth of GaAs nanowires (NWs) with controllable number density and diameter, which consists of the combination between droplet epitaxy (DE) and self-assisted NW growth. In our method, GaAs islands are initially formed on Si(111) by DE and, subsequently, GaAs NWs are selectively grown on their top facet, which acts as a nucleation site. By DE, we can successfully tailor the number density and diameter of the template of initial GaAs islands and the same degree of control is transferred to the final GaAs NWs. We show how, by a suitable choice of V/III flux ratio, a single NW can be accommodated on top of each GaAs base island. By transmission electron microscopy, as well as cathodo- and photoluminescence spectroscopy, we confirmed the high structural and optical quality of GaAs NWs grown by our method. We believe that this combined approach can be more generally applied to the fabrication of different homo- or heteroepitaxial NWs, nucleated on the top of predefined islands obtained by DE. KEYWORDS: Nanowires, droplet epitaxy, self-organization, nanostructures, silicon, compound semiconductors

S

situ techniques and self-organization phenomena. In order to achieve this level of control, we make use of droplet epitaxy (DE), a flexible growth method for the fabrication of III−V semiconductor nanostructures, originally proposed by Koguchi in 1991.13 The DE technique relies on the controlled formation of group-III element droplets, which are then crystallized into III−V nanostructures by means of an intense flux of a group-V element. By this method, several kinds of simple and complex GaAs nanostructures have already been demonstrated.14−19 In addition, DE allows for the independent tuning of droplet, and thus nanostructure number density and diameter, through a careful choice of the initial deposition conditions of the groupIII element.20 In our work, we make use of this degree of freedom and form GaAs islands on Si(111) with controllable number density and diameter, which then serve as nucleation sites for GaAs NWs. These islands can therefore be considered as a template, that is completely fabricated in situ without the need of any pregrowth effort, such as an efficient mask for the following NW growth. Under suitable growth conditions, a GaAs NW can be grown on top of each GaAs island, similarly to what could be obtained by a SAG process. All samples were grown in a molecular beam epitaxy (MBE) system on Si(111) substrates. The formation of Ga droplet and GaAs island was performed on both oxide-free wafers (dipped in HF right before the insertion into the MBE chamber) and on substrates covered by the native oxide. In the first step of our

emiconductor nanowires are considered promising building blocks for a new generation of devices,1−4 as well as ideal platforms for advanced investigations in fundamental physics.5−7 In particular, GaAs NWs grown without the use of gold as a catalyst8,9 recently attracted a lot of attention, as they may pave the way for the monolithic integration of III−V materials with the mature silicon technology. However, in this self-assisted growth mode, tuning NW number density and diameter still remains a challenge. Nevertheless, achieving such kind of control over the fabrication process would surely be desirable for both fundamental studies and practical applications. According to the present understanding of self-assisted NW nucleation, Ga droplets are spontaneously formed on the substrate surface, that is covered by a thin silicon oxide layer, and induce GaAs NW growth.10 The epitaxial relationship with the substrate is ensured through nanometer-sized pinholes, which are randomly created in the silicon oxide layer and permit the contact between NWs and substrate.8 Although a clear connection between NW and pinhole geometrical features has not been demonstrated yet, the nucleation of GaAs NWs seems to be related to the presence of these random imperfections in the oxide layer. Moreover, pinhole formation, Ga droplet nucleation, and GaAs growth are competing processes and controlling the number density and diameter of self-assisted GaAs NWs is very difficult. Naturally, substrate prepatterning, followed by selective area growth (SAG), could also be used,11,12 but it requires a considerable technological effort in terms of time, expertise, and equipment. Instead, here we propose a novel approach to overcome this limitation and open the possibility for tuning the number density and diameter of self-assisted GaAs NWs, all based on in © XXXX American Chemical Society

Received: April 19, 2013 Revised: July 3, 2013

A

dx.doi.org/10.1021/nl401404w | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

the Ga droplet ensemble were statistically derived from AFM images, collected in tapping mode. In more detail, the droplet diameter and number density were 110 nm and 1.8 × 108 cm−2 (Figure 1a), 106 nm and 2.4 × 107 cm−2 (Figure 1b), 202 nm and 1.9 × 108 cm−2 (Figure 1c), and 212 nm and 2.6 × 107 cm−2 (Figure 1d), with a droplet size dispersion always around 10%, as shown in Figure 1e. A representative histogram of the droplet size distribution (the data belong to the sample in Figure 1c) is here reported, with volumes normalized to the most probable value, together with a Gaussian fit of the distribution (in red). Similarly to what has already been described for Ga droplet formation on GaAs(001)20 and Si(001)21 substrates, we can reliably control the droplet number density over a wide range on Si(111) also, by a suitable choice of the substrate temperature: for higher temperatures, a lower number density is obtained and vice versa. In our case, droplets were formed at a substrate temperature of 340 °C (Figure 1a,c) and 460 °C (Figure 1b,d). To quantitatively summarize this trend, we show the Arrhenius plot of gallium droplet number density as a function of temperature in Figure 1f. This phenomenon can be understood in terms of an increase in Ga adatom diffusion length for higher substrate temperatures. Under these conditions, an impinging Ga adatom will have enough mobility to reach an already existing cluster and get incorporated there; on the contrary, for lower substrate temperature, the adatom diffusion length will be reduced and, consequently, more droplets will nucleate on the substrate surface. If the same deposition conditions are employed but a larger amount of Ga is supplied, droplets with the same number density but larger diameter can be obtained. This is clearly demonstrated by the comparison between the samples shown in Figure 1a,c and b,d, respectively, where the amount of supplied Ga in each case was 5, 20 ML and 0.5, 2 ML. In this study, we explored for the droplet number density the range 106−109 cm−2 and for the droplet diameter 10−450 nm. We expect that it is possible to expand these ranges even further, by using different sets of Ga deposition conditions. In addition, the variation in droplet number density over the whole substrate (2 in.) was below ±5%. Qualitatively, similar results and the same degree of control can be obtained when using Si(111) substrates covered by the native oxide. After the formation of the desired Ga droplet ensemble, GaAs islands were obtained by arsenic irradiation (As flux around 1 × 10−5 mbar for 10 min). In DE, during the arsenization step, each Ga droplet is directly transformed into a GaAs island,14,15 and therefore, the same degree of control gained over Ga droplets is transferred to crystallized GaAs islands. This means that we can create a template made of GaAs stems with the desired number density and diameter. Indeed, as previously reported,22,23 no big change in diameter occurs after the crystallization of liquid Ga into GaAs, as the Ga droplet perimeter also sets the GaAs island edges, due to its high reactivity. The expected increase in volume caused by the arsenization of the droplet is mainly accommodated by a change in shape and by an increase of the aspect ratio from the droplet to the island. As examples, we show different island ensembles with diameters and number densities of around 425 nm and 1.8 × 107 cm−2 (Figure 2a), 450 nm and 1.0 × 108 cm−2 (Figure 2b), and 250 nm and 4.0 × 107 cm−2 (Figure 2c). Clearly, the island number density and diameter can be independently controlled, as GaAs stems with similar diameter but different number density (Figure 2a and b) or with similar

approach, Ga droplet formation was achieved by supplying Ga at a given substrate temperature, without simultaneous supply of As, which would disturb the controlled nucleation of Ga droplets. As shown by the atomic force microscopy (AFM) images of four representative samples in Figure 1, droplets of

Figure 1. AFM images of Ga droplets on oxide-free Si(111) having different number densities (a) 1.8 × 108 cm−2, (b) 2.4 × 107 cm−2, (c) 1.9 × 108 cm−2, and (d) 2.6 × 107 cm−2 and diameters (a) 110 nm, (b) 106 nm, (c) 202 nm, and (d) 212 nm. Scale bars are 1 μm. (e) Histogram of the droplet size distribution for the sample reported in part c, with volumes normalized to the most probable value; the red curve shows the Gaussian fit of the distribution. (f) Arrhenius plot of Ga droplet number density as a function of substrate temperature.

different size and number density can be formed on oxide-free Si substrates and these two parameters can be independently controlled. Indeed, droplets of different number density but nearly identical diameter can be obtained (Figure 1a,b and c,d, respectively), and also different diameters at very similar number densities can be achieved (Figure 1a,c and b,d, respectively). The average diameter and number density of B

dx.doi.org/10.1021/nl401404w | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

etched with HCl. In this way, we could remove the native oxide from GaAs but not from Si, reducing the amount of parasitic growth on the substrate. We fabricated a series of samples starting with predefined islands obtained under identical conditions and varied the V/III flux ratio during the subsequent growth of GaAs. In particular, the used template, already shown in Figure 2c and d, possessed a GaAs island number density and diameter of 4.0 × 107 cm−2 and 250 nm, respectively. In every experiment, the substrate was kept at the typical GaAs NW growth temperature of 580 °C, the Ga flux at 100 nm/h, and the growth time equal to 15 min and the As flux was systematically varied in order to change the corresponding V/ III flux ratio between 1 and 8. Figure 3 shows the results of this growth study. When small V/III ratios (1 or 2) were employed (Figure 3a,b), mainly large Ga droplets covered the pre-existing GaAs islands, without the formation of any NW. On the other hand, for intermediate V/III ratios (3), many NWs formed on top of the base islands and, in addition, a fairly small number of parasitic islands grew directly on the substrate between the original islands (Figure 3c). From now on, this sample will be referred to as sample A. Finally, when large V/III flux ratios (4 or 8) were used in this step (Figure 3d,e), some NWs could grow on top of the original GaAs islands, but the substrate became more and more covered by parasitic islands. NWs obtained at the end of our procedure are morphologically similar to MBE-grown GaAs NWs, fabricated by the selfassisted vapor−liquid−solid (VLS) growth, as already reported by several groups.8,9,24−27 In addition, we observe a clear enlargement at the base, which we relate to the presence of the initial GaAs island, suggesting that NWs grew on top of the predefined stems. Moreover, we see the following: (i) In sample A, no NWs grew directly on the substrate; therefore, every wire that we observed at the end of the growth was accommodated on top of an initial GaAs island base. (ii) We compared the number density of the initial GaAs island template and the final NWs and obtained in both cases the same value (4.0 × 107 cm−2). This means that, under suitable conditions, the yield for NW growth on the in situ formed island template can be as high as 100%. In this particular case, NWs with a diameter of around 120 nm and a length (measured from the island top to the droplet bottom) of around 1 μm were fabricated. The presented results provide some insights about the NW growth kinetics, as very different morphologies were obtained

Figure 2. Plan-view (a−c) and cross-sectional (d) SEM images of GaAs islands on Si(111) after crystallization of Ga droplets by As flux. The diameter and number density were (a) 1.8 × 107 cm−2 and 425 nm, (b) 1.0 × 108 cm−2 and 450 nm, and (c) 4.0 × 107 cm−2 and 250 nm, respectively. The scale bars are 3 μm in parts a−c and 100 nm in the inset of part c and in part d.

number density but different diameter (Figure 2a and c) can be fabricated. Interestingly, we observe that, after the crystallization, these GaAs islands grown on Si(111) by droplet epitaxy show typical NW-like features. Indeed, a perfect hexagonal cross section is typically observed, when imaging GaAs islands from the top (inset of Figure 2c) and well-defined vertical side walls are found, when looking at the island from the side (Figure 2d). Additionally, a flat facet is clearly developed at the top of the islands, locally providing a GaAs(111) surface, where the following GaAs NW nucleation will actually take place. Next, we investigated the possibility of using the islands formed by DE as nucleation sites for NW growth. After GaAs base islands were formed on bare Si(111), the samples were taken out from the growth chamber, oxidized in air, and finally

Figure 3. Tilted (25°) SEM images of samples after the GaAs growth on predefined islands at different V/III ratios. (a) V/III = 1, (b) V/III = 2, (c) V/III = 3 (sample A), (d) V/III = 4, (e) V/III = 8. Scale bars are 1 μm. C

dx.doi.org/10.1021/nl401404w | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

by modifying the growth conditions. In order to selectively obtain the homoepitaxial growth of GaAs on the GaAs stems but not on the surrounding native Si oxide, Ga adatoms must possess enough surface mobility to reach an island. For a constant substrate temperature, this condition will be fulfilled at lower V/III flux ratios, but at the same time, NW growth is possible only if, locally, a sufficient amount of As atoms is present; otherwise, a situation similar to what is shown in Figure 3a will occur. Naturally, an optimization of the growth conditions between the two extreme cases of too low V/III ratio leading to large droplets and of too high V/III ratio leading to very pronounced parasitic growth is necessary for each arrangement of the initial GaAs island template. Here, we have shown that the highest yield for VLS NW growth was obtained for V/III ratio equal to 3, which means nominally Asrich. However, Ga droplets could clearly form under these conditions, confirming that what matters for NW growth is the local V/III flux ratio. Another important factor of NW growth kinetics is the rate-limiting species. For self-assisted GaAs NWs, it is known that the growth rate depends linearly on As flux,10 meaning that arsenic is the rate limiting species, in contrast to the standard GaAs planar MBE growth.28 In our case, we can compare the NW growth rate (4 μm/h) in the case of V/III ratio equal to 3 in our method (sample A) to the growth rate (1.3 μm/h) we observed for standard self-assisted GaAs NWs grown under V/III ratio equal to 1 without the mediation of DE (as shown in the Supporting Information, section S1). Since for these two samples the ratio between growth rates matches the ratio between As fluxes, we can conclude that the growth of GaAs NWs on top of islands fabricated by DE is governed by the same growth kinetics found in the self-assisted method. In order to demonstrate control over number density and diameter for the final GaAs NWs, we present sample B. The aim for this sample was to grow a NW ensemble with strongly different features than in sample A. As shown in Figure 4, this time the final NWs possess a higher number density (3 × 108 cm−2) and a much thinner diameter of around 40 nm. NWs with such a small diameter are normally not obtained when using the standard Ga-assisted method. Again, a Ga droplet is present on their top, suggesting identical growth kinetics, and some parasitic growth is also observed on the substrate surface. In this case, in order to obtain the highest NW yield, the V/III ratio was kept equal to 1. Again, we found a clear tendency for NWs to be formed on top of predefined GaAs islands and not directly on the substrate surface, meaning that the homoepitaxial growth of NWs on top of the base islands is energetically favorable. However, some NWs seem to grow directly on the silicon substrate, as they do not show a clear base at their very bottom. Interestingly, for both sample A and sample B, very similar values are obtained when calculating the ratio between the diameter of the final NW and the diameter of the original GaAs base island. For sample A, the NW diameter was 120 nm and the island diameter 250 nm, while, for sample B, the diameters were 40 and 90 nm, respectively, therefore leading to a ratio of 0.48 for sample A and 0.44 for sample B. This similarity is probably caused by a proportionality in the reduced size of the flat top facet of the base island compared to the island diameter (see Figure 2d). This observation suggests that the diameter of the NWs we can obtain is about half of the diameter of the base islands. To assess the microstructural properties of NWs grown by our novel approach, transmission electron microscopy (TEM)

Figure 4. (a) Plan-view and (b) tilted (25°) SEM images of NW sample B with high number density (3 × 108 cm−2) and small diameter (40 nm). Scale bars are 500 nm.

was performed on the two presented samples (A and B) and is reported in Figure 5. Two structures from sample A, where the typical geometrical features of GaAs NWs on top of GaAs islands can be clearly recognized, are shown in Figure 5a. Moving from the bottom to the top, we first observe the base grown by DE, which can be easily distinguished because of the larger diameter (250 nm); second, we find a region, still belonging to the island at the bottom, where the diameter reduces (as already reported in Figure 2d); then, we observe the GaAs NWs with regular diameter (120 nm); and finally, we clearly see the Ga droplet at the NW tip, as further confirmed in the Supporting Information (section S2). Very interesting for our method is the study of the interface region between the island grown by DE and the NW, which is expected to be immediately above the area where the structure diameter reduces from the base island to the NW. A representative highresolution TEM (HR-TEM) image of this interface region, taken from sample B, is shown in Figure 5b. The GaAs stems adopt the zincblende (ZB) crystal structure and several planar defects, stacking faults (SFs), and nanotwins can normally be found in their top part, in correspondence with the diameter reduction. On the contrary, defects were not detected in the junction region and the NWs grow on top of the stems with a perfect epitaxial quality. As a matter of fact, the actual interface cannot be identified in such micrographs. These observations D

dx.doi.org/10.1021/nl401404w | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

size.33 This result is very important, as such defects can reduce the performance of NW-based devices34 and induce very complex carrier recombination mechanisms, as already reported several times.30,35−38 Finally, we have prepared a dedicated sample (sample C) to investigate the optical properties of GaAs NWs grown by this novel approach, capped by an (Al,Ga)As shell (33% nominal aluminum content), in order to reduce the nonradiative recombination at the free GaAs surface.27,39,40 Here, NWs possess a diameter of around 130 nm, which makes them similar to the already described sample A, and a number density of around 2 × 107 cm−2. In order to distinguish between luminescence from NWs and parasitic growth, we characterized the sample by cathodoluminescence (CL) spectroscopy. In Figure 6a, the spatially resolved CL spectra of a single structure

Figure 5. TEM micrographs of GaAs NWs grown on DE islands. (a) Overview of two complete NW structures from sample A, (b) HRTEM image of the interface region between the island grown by DE (bottom right) and the NW (top left) in a structure from sample B, (c) large diameter (120 nm) NW from sample A, with a high defect density, (d) small diameter (40 nm) NW from sample B, with three SFs, marked by arrows. In the insets of parts c and d, the corresponding SAD patterns are shown.

Figure 6. (a) CL 2D plot of a single, as-grown GaAs/(Al,Ga)As NW at 10 K with the emission intensity color-coded on a logarithmic scale. Monochromatic CL superimposed to the respective SEM image, corresponding to (b) 820 nm and (c) 790 nm emission wavelengths. (d) Room temperature μ-PL spectrum of the same sample.

confirm the excellent epitaxial quality and the atomic smoothness of the interface between NW and base. Moreover, the first few hundreds of nanometers of the NW normally do not contain any SF at all. The high phase purity of the lower NW segment seems to be a general feature, as we found the same also in sample A. Therefore, we conclude that DE islands can be used as an efficient template for the overgrowth of high quality semiconductors. In Figure 5c and d, we show a more detailed TEM investigation of the NW upper segment for samples A and B, respectively. In both cases, as visible in the corresponding selective area diffraction (SAD) patterns in the insets, GaAs NWs crystallize in the ZB phase, which is typically observed in the self-assisted growth mode,8,9,25 but we found a marked difference in the two samples regarding the defect density in the NW top part. Indeed, in the case of sample A, we detected SFs, twin planes, and defective regions, which is again a very common observation for self-assisted GaAs NWs. 29−31 However, these defects are normally not contained in the NW lower segment, in agreement with a previous report.32 Twins are clearly visible in the corresponding SAD pattern, which also appears more streaky along the ⟨111⟩ direction, because of the structural defects. On the contrary, in NWs from sample B, we always found very few and frequently even no SFs and, therefore, a higher crystal purity along the whole NW length. As an example, Figure 5d shows a NW with three SFs, marked by arrows. The better crystal quality of thinner NWs is also demonstrated by the SAD pattern shown in the inset. This difference in defect density for samples A and B seems to be related to their different diameter. The much lower defect density for the thinner NWs with 40 nm diameter is in agreement with what was found previously in NWs of similar

measured at 10 K are presented as a 2D plot with the intensity color coded. For all the NWs investigated by CL, we detected two main bands, centered at around 820 and 790 nm. The strong emission peak at 820 nm originates from the lower segment of these structures, while the weaker high-energy emission stems from their upper part. This result is confirmed more illustratively in Figure 6b and c, showing the monochromatic CL images for the two characteristic wavelengths superimposed on the corresponding SEM image of the same NW measured in Figure 6a. As already mentioned, by TEM, we found the lower part of the NWs to have a very low defect density and we can therefore assign the emission at around 820 nm to carrier recombination in ZB GaAs, in good agreement with the literature value.41 On the contrary, as was shown in Figure 5c for a structure with similar diameter, the NW top part contains many structural defects, namely, alternating ZB and WZ segments, which may induce radiative recombination at energies exceeding the ZB band gap,35,36,38 as found by CL in Figure 6a and c. We believe this effect to be the origin of the emission peak at around 790 nm. No emission from the parasitic growth was instead detected by CL, allowing us to safely characterize this sample also by μ-photoluminescence (PL). For the excitation, we used a HeNe laser operating at 632.8 nm, focused on a spot with a diameter of around 1 μm, allowing us to measure, on average, less than 1 NW at time. Measurements were carried out on several spots per sample, and very similar spectra were acquired. In Figure 6d, we show a room temperature μ-PL spectrum which is dominated by an intense emission centered at around 870 nm, E

dx.doi.org/10.1021/nl401404w | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

again in good agreement with the energy gap of ZB GaAs,41 with a high-energy tail which may again be caused by the structural disorder. The presence of a clear PL emission even at room temperature points toward the high optical quality of NWs grown by the mediation of DE. In conclusion, we presented a new method for the growth of self-assisted GaAs NWs mediated by DE. The key idea behind this approach is the use of GaAs islands, formed by the crystallization of Ga droplets, as nucleation sites for NW growth. These base islands are fabricated in situ with highly controllable diameter and number density and do not require any pregrowth effort. We have shown how it is possible to achieve, under optimized conditions, a perfect positioning of a single NW on top of each initial base island. In this way, the same degree of control can be transferred from the stems to the NW geometrical features, making our method suitable for the fabrication of GaAs NWs with controllable number density and diameter. By TEM, we found that a perfect, defect-free interface connects the initial base islands and the NWs on their top, meaning that the growth proceeds maintaining a high epitaxial quality. Furthermore, we detected a much lower SF density in thinner NWs (diameter of 40 nm), which seems to suggest a better crystal phase purity for NWs having smaller diameter. Moreover, CL and μ-PL investigations on capped GaAs NWs revealed the high optical quality of NWs grown by this method. Finally, it is worth pointing out that the presented use of DE islands as templates for overgrowth, similarly to the growth on prepatterned substrates, is not limited to the homoepitaxial case of GaAs on GaAs. Indeed, we believe that our method can surely be extended to other homo- and heteroepitaxial systems, as, for example, InAs on GaAs.



(4) Cui, Y.; Zhong, Z.; Wang, D.; Wang, W. U.; Lieber, C. M. Nano Lett. 2003, 3, 149−152. (5) Mourik, V.; Zuo, K.; Frolov, S. M.; Plissard, S. R.; Bakkers, E. P. a. M.; Kouwenhoven, L. P. Science 2012, 13, 1003. (6) Fasth, C.; Fuhrer, A.; Samuelson, L.; Golovach, V. N.; Loss, D. Phys. Rev. Lett. 2007, 98, 266801. (7) Cho, C.-H.; Aspetti, C. O.; Turk, M. E.; Kikkawa, J. M.; Nam, S.W.; Agarwal, R. Nat. Mater. 2011, 10, 669−675. (8) Fontcuberta i Morral, A.; Colombo, C.; Abstreiter, G.; Arbiol, J.; Morante, J. R. Appl. Phys. Lett. 2008, 92, 63112. (9) Jabeen, F.; Grillo, V.; Rubini, S.; Martelli, F. Nanotechnology 2008, 19, 275711. (10) Colombo, C.; Spirkoska, D.; Frimmer, M.; Abstreiter, G.; Fontcuberta i Morral, A. Phys. Rev. B 2008, 77, 155326. (11) Ikejiri, K.; Noborisaka, J.; Hara, S.; Motohisa, J.; Fukui, T. J. Cryst. Growth 2007, 298, 616. (12) Bauer, B.; Rudolph, A.; Soda, M.; Fontcuberta i Morral, A.; Zweck, J.; Schuh, D.; Reiger, E. Nanotechnology 2010, 21, 435601. (13) Koguchi, N.; Takahashi, S.; Chikyow, T. J. Cryst. Growth 1991, 111, 688. (14) Koguchi, N.; Ishige, K. Jpn. J. Appl. Phys. 1993, 32, 2052. (15) Watanabe, K.; Koguchi, N.; Gotoh, Y. Jpn. J. Appl. Phys. 2000, 39, L79. (16) Mano, T.; Kuroda, T.; Sanguinetti, S.; Ochiai, T.; Tateno, T.; Kim, T.; Noda, T.; Kawabe, M.; Sakoda, K.; Kido, G.; Koguchi, N. Nano Lett. 2005, 5, 425. (17) Huang, S.; Niu, Z.; Fang, Z.; Ni, H.; Gong, Z.; Xia, J. Appl. Phys. Lett. 2006, 89, 031921. (18) Wang, Z. M.; Liang, B. L.; Sablon, K. A.; Salamo, G. J. Appl. Phys. Lett. 2007, 90, 113120. (19) Somaschini, C.; Bietti, S.; Koguchi, N.; Sanguinetti, S. Nano Lett. 2009, 9, 3419. (20) Heyn, C.; Stemmann, A.; Schramm, A.; Welsch, H.; Hansen, W.; Nemcsics, A. Phys. Rev. B 2007, 76, 075317. (21) Somaschini, C.; Bietti, S.; Sanguinetti, S.; Koguchi, N.; Montalenti, F.; Frigeri, C. Appl. Phys. Lett. 2010, 97, 053101. (22) Mano, T.; Koguchi, N. J. Cryst. Growth 2005, 278, 108. (23) Bietti, S.; Somaschini, C.; Sanguinetti, S. Nanotechnology 2013, 24, 205603. (24) Paek, J. H.; Nishiwaki, T.; Yamaguchi, M.; Sawaki, N. Phys. Status Solidi C 2009, 6, 1436−1440. (25) Cirlin, G. E.; Dubrovskii, V. G.; Samsonenko, Y. B.; Bouravleuv, A. D.; Durose, K.; Proskuryakov, Y. Y.; Mendes, B.; Bowen, L.; Kaliteevski, M. A.; Abram, R. A.; Zeze, D. Phys. Rev. B 2010, 82, 35302. (26) Plissard, S.; Dick, K. A.; Wallart, X.; Caroff, P. Appl. Phys. Lett. 2010, 96, 121901. (27) Breuer, S.; Pfüller, C.; Flissikowski, T.; Brandt, O.; Grahn, H. T.; Geelhaar, L.; Riechert, H. Nano Lett. 2011, 11, 1276−1279. (28) Neave, J. H.; Dobson, P. J.; Joyce, B. A.; Zhang, J. Appl. Phys. Lett. 1985, 47, 100. (29) Yu, X.; Wang, H.; Lu, J.; Zhao, J.; Misuraca, J.; Xiong, P.; von Molnár, S. Nano Lett. 2012, 12, 5436−5442. (30) Spirkoska, D.; et al. Phys. Rev. B 2009, 80, 245325. (31) Rudolph, D.; Hertenberger, S.; Bolte, S.; Paosangthong, W.; Spirkoska, D.; Doeblinger, M.; Bichler, M.; Finley, J. J.; Abstreiter, G.; Koblmüller, G. Nano Lett. 2011, 11, 3848−3854. (32) Plissard, S.; Dick, K. A.; Larrieu, G.; Godey, S.; Addad, A.; Wallart, X.; Caroff, P. Nanotechnology 2010, 21, 385602. (33) Krogstrup, P.; Popovitz-Biro, R.; Johnson, E.; Madsen, M. H.; Nygå rd, J.; Shtrikman, H. Nano Lett. 2010, 10, 4475−4482. (34) Thelander, C.; Caroff, P.; Plissard, S.; Dey, A. W.; Dick, K. A. Nano Lett. 2011, 11, 2424−2429. (35) Hoang, T. B.; Moses, a. F.; Zhou, H. L.; Dheeraj, D. L.; Fimland, B. O.; Weman, H. Appl. Phys. Lett. 2009, 94, 133105. (36) Jancu, J.-M.; Gauthron, K.; Largeau, L.; Patriarche, G.; Harmand, J.-C.; Voisin, P. Appl. Phys. Lett. 2010, 97, 041910. (37) Heiss, M.; Conesa-Boj, S.; Ren, J.; Tseng, H.-H.; Gali, A.; Rudolph, A.; Uccelli, E.; Peiró, F.; Morante, J. R.; Schuh, D.; Reiger,

ASSOCIATED CONTENT

S Supporting Information *

Figures showing large-area and single-structure SEM images of GaAs NWs grown with and without the mediation of droplet epitaxy and SEM images of two pieces of sample A before and after etching by HCl. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors would like to thank Claudia Herrmann for the maintenance of the MBE system, Anne-Kathrin Bluhm for the SEM images, Doreen Steffen for the preparation of TEM specimens, and Jonas Lähnemann for the critical reading of the manuscript. This work was supported by the Deutsche Forschungsgemeinschaft (DFG) under grant Ge2224/2.



REFERENCES

(1) Huang, M. H.; Mao, S.; Feick, H.; Yan, H.; Wu, Y.; Kind, H.; Weber, E.; Russo, R.; Yang, P. Science 2001, 292, 1897−1899. (2) Duan, X.; Huang, Y.; Cui, Y.; Wang, J.; Lieber, C. M. Nature 2001, 409, 66−69. (3) Bjork, M. T.; Ohlsson, B. J.; Thelander, C.; Persson, A. I.; Deppert, K.; Wallenberg, L. R.; Samuelson, L. Appl. Phys. Lett. 2002, 81, 4458−4460. F

dx.doi.org/10.1021/nl401404w | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

E.; Kaxiras, E.; Arbiol, J.; Fontcuberta i Morral, A. Phys. Rev. B 2011, 83, 45303. (38) Jahn, U.; Lähnemann, J.; Pfüller, C.; Brandt, O.; Breuer, S.; Jenichen, B.; Ramsteiner, M.; Geelhaar, L.; Riechert, H. Phys. Rev. B 2012, 85, 045323. (39) Demichel, O.; Heiss, M.; Bleuse, J.; Mariette, H.; Fontcuberta i Morral, A. Appl. Phys. Lett. 2010, 97, 201907. (40) Parkinson, P.; Joyce, H. J.; Gao, Q.; Tan, H. H.; Zhang, X.; Zou, J.; Jagadish, C.; Herz, L. M.; Johnston, M. B. Nano Lett. 2009, 9, 3349−3353. (41) Grilli, E.; Guzzi, M.; Zamboni, R.; Pavesi, L. Phys. Rev. B 1992, 45, 1638−1644.

G

dx.doi.org/10.1021/nl401404w | Nano Lett. XXXX, XXX, XXX−XXX