Fast Growth Synthesis of GaAs Nanowires with Exceptional Length

Apr 12, 2010 - UMR 6602, LASMEA, F-63177 Aubie`re, France, § Université du Sud Toulon-Var, IM2NP, Bât. R, B.P. 20132,. 83957 La Garde Cedex, France...
2 downloads 0 Views 335KB Size
Download PDF
pubs.acs.org/NanoLett

Fast Growth Synthesis of GaAs Nanowires with Exceptional Length M. R. Ramdani,†,‡ E. Gil,*,†,‡ Ch. Leroux,§,| Y. Andre´,†,‡ A. Trassoudaine,†,‡ D. Castelluci,†,‡ L. Bideux,†,‡ G. Monier,†,‡ C. Robert-Goumet,†,‡ and R. Kupka†,‡ †

Clermont Universite´, Universite´ Blaise Pascal, LASMEA, BP 10448, F-63000 Clermont-Ferrand, France, ‡ CNRS, UMR 6602, LASMEA, F-63177 Aubie`re, France, § Universite´ du Sud Toulon-Var, IM2NP, Baˆt. R, B.P. 20132, 83957 La Garde Cedex, France, and | CNRS, UMR 6242, 83957 La Garde Cedex, France ABSTRACT We report the first synthesis of GaAs nanowires (NWs) by Au-assisted vapor-liquid-solid (VLS) growth in the novel hydride vapor phase epitaxy (HVPE) environment. Forty micrometer long rodlike 〈111〉 monocrystalline GaAs nanowires exhibiting a cubic zinc blende structure were grown in 15 min with a mean density of 106 cm-2. The synthesis of such long figures in such a short duration could be explained by the growth physics of near-equilibrium HVPE. VLS-HVPE is mainly based on solidification after direct and continuous feeding of the arsenious and GaCl growth precursors through the Au-Ga liquid catalyst. Fast solidification (170 µm/h) is then assisted by the high decomposition frequency of GaCl. This predominant feeding through the liquid-solid interface with no mass and kinetic hindrance favors axial rather than radial growth, leading to twin-free nanowires with a constant cylinder shape over unusual length. The achievement of GaAs NWs several tens of micrometers long showing a high surface to volume ratio may open the field of III-V wires, as already addressed with ultralong Si nanowires. KEYWORDS GaAs, nanowires, hydride vapor phase epitaxy

N

anowires (NWs) are confirmed nanocomponent candidates for optoelectronic devices1-3 or sensor and transistor applications.4,5 Controlling the wire shaping is the first challenge when addressing the synthesis of NWs; one requires uniform morphology from bottom to top of the wires and homogeneous crystalline structure. There is also a need for dense arrays of position-controlled nano-objects. Nanowires are synthesized by a wide range of methods: top-down photolitography,6 and bottom-up catalyst-free selective area (SA) growth or catalyst-assisted vapor-liquid-solid (VLS) growth. Whatever the route, either selective condensation or VLS solidification, the growth techniques involved in the synthesis of III (Ga, In) -V (P, As) nanowires are molecular beam epitaxy (MBE) and metal organic vapor phase epitaxy (MOVPE). MOVPE benefits from a worldwide presence for III-V issues. It is often considered as the most promising for the industrial mass manufacture of semiconductor nanowire arrays. As a result, remarkable arrays of highly uniform size-, shape-, and position-controlled GaAs and InAs hexagonal nanopillars were grown by SA-MOVPE on patterned substrates.7,8 One has to note that nanopillars grown by SA-MOVPE exhibit limited lengths (lower than 6 µm, depending on the nanopillar diameter) for 60 min growth. The process route involves a preliminary patterning step of the substrate on a nanometer scale, which provides an accurate control of the NW position and diameter (from 50 to 200 nm) through circular openings in the

dielectric film masking the substrate. The nanopillar growth rate is indeed greater than standard planar growth for otherwise identical experimental conditions but still remains weak, close to 5 µm/h. A major advantage of selective area epitaxy is that planar growth is inhibited by the dielectric film. The alternative bottom-up approach for the synthesis of NWs is the catalyst-assisted vapor-liquid-solid growth as introduced by Wagner and Ellis in 1964 to explain the epitaxial growth of micrometer-sized Si whiskers.9 The second fundamental work was provided by Hiruma et al. who demonstrated the feasibility of the VLS growth of III-V nanowhiskers.10 In VLS growth, a suitable metal catalyst alloys with the metal atoms of the III-V surface introducing a local liquid-solid interface. The constituent materials are supplied from the vapor phase and incorporated into the metal catalyst, which acts as a seed for the NW growth. In most cases, GaAs NWs are grown from Au catalytic particles. Semiconductor NW VLS growth is carried out by MBE,11-14 chemical beam epitaxy (CBE),15 electron beam epitaxy (EBE),16 and MOVPE.17-22 The metallic clusters play a very important role in determining the shape and position of nanowires. Ohlsson et al. used size-selected catalytic Au aerosol particles that were manipulated by means of atomic force microscopy in order to nucleate nanowhiskers in specific positions.23 Patterning of the catalyst on the substrate surface is also carried out by electron beam lithography,24 or nanoimprint lithography,25 followed by catalyst evaporation and lift-off. In this letter, we report the first growth of GaAs NWs by hydride vapor phase epitaxy (HVPE). HVPE has been known

* To whom correspondence should be addressed. E-mail: Evelyne.GIL@ lasmea.univ-bpclermont.fr. Tel: 33 4 73 40 73 44. Fax: 33 4 73 40 73 40. Received for review: 02/15/2010 Published on Web: 04/12/2010 © 2010 American Chemical Society

1836

DOI: 10.1021/nl100557d | Nano Lett. 2010, 10, 1836–1841

for a while to be a near equilibrium condensation process. The growth of GaAs by HVPE involves gaseous GaCl molecules synthesized inside the reactor by the use of HCl flow reacting with a Ga liquid source. Element V is transported as arsine (AsH3) which is completely decomposed into (As2, As4) molecules considered to be at equilibrium by entering the hot wall reactor.26 The main feature of HVPE is the use of chloride molecules as element III growth precursors; the dechlorination frequency is high enough so that there is no kinetic delay and equilibrium conditions are quickly reached.27-29 A wide range of condensation growth rates, from 3 to 100 µm/ h, can then be set due to immediate reactivity to an increase (or decrease) of the vapor phase supersaturation. As a matter of fact, the HVPE reactor geometry and the experimental parameters can be chosen in such a way that gas-phase mass transport is not a limiting factor. Growth is then governed by surface kinetics, as demonstrated by the strong temperature dependence of the GaAs growth rate under optimized conditions.28 Among the surface kinetics (adsorption and decomposition of growth precursors, desorption of byproduct, and diffusion of ad-species), the GaCl growth precursor decomposition kinetics appears to be the slowest, so that it is considered as the actual limiting rate for HVPE growth.28 Being developed in the early 60s for the growth of the first III-V epilayers, HVPE has been neglected in favor of MOVPE and MBE as the need for low dimensionality heterostructures has been increasing. HVPE was reintroduced with the revival of nitride compounds. Here, HVPE offers fast growth properties (100 µm/h for GaN epitaxy), which are adapted to the synthesis of GaN free-standing quasi-substrates. HVPE is the least polluting and cheapest VPE process. It makes use of few materials: a vector gas flow rate of 3 L/min as opposed to 6 L/min for MOVPE; 10 SCCM of V growth precursors (usually arsine for As) compared to 300 SCCM needed for MOVPE. The aim is not to oppose HVPE to established MOVPE and MBE, but to propose it as a complementary approach for identified synthesis objectives, where HVPE may appear as more effective and more economical and less polluting. Very interestingly, Joyce et al. used rapid growth rate in VLS-MOVPE of III-V nanowires to synthesize less tapered, planar crystallographic defect-free nanowires.30 One should promote untapered morphologies for laser applications where uniform diameters are expected to enhance the NW performance as a resonant cavity. Use of rapid growth was addressed for the IV compounds with the synthesis of silicon and germanium nanowires with lengths up to 30 µm in a pioneer work.31 Highly uniform millimeter-long Si NWs were successfully grown at a mean rate of 31 µm/min and used as building blocks to fabricate field effect transistor arrays consisting of some 100 devices per nanowire.32 Nanowires a few tens of micrometers long provide larger surfaces to be functionalized, which could be of great interest for sensor and detection applications.32,33 Because of the rapid decomposition of growth precursors, one could expect HVPE to be © 2010 American Chemical Society

an effective alternative tool for the synthesis of III-V alloy nanowires with exceptional length at an acceptable growth rate, provided the control of the morphology and crystalline homogeneity of the nano-objects. We have therefore investigated the synthesis of HVPE-GaAs NWs formed by the VLS growth mode, assisted by a nonpatterned Au catalytic layer deposited prior to GaAs growth. For this preliminary study, we have chosen nonoptimum conditions like early MOVPE and MBE attempts, random Au distribution on the surface and (001) oriented GaAs substrates. The aim was to assess (i) whether VLS could take place in a HVPE environment, in particular at high temperature that is needed to ensure the stability of the chloride precursors, and (ii) the competition between VLS expected to take place in the 〈111〉 direction and standard condensation planar growth on (001). GaAs VLS-HVPE growth was carried out on 4° misoriented vicinal (001) GaAs substrates in a hot wall horizontal reactor kept at atmospheric pressure (AP-HVPE). The substrates were rinsed, treated with diluted hydrochloric acid (HCl) to remove the native oxide, and degreased. A uniform 2 nm thick Au film was deposited on the surface of each substrate in an ultrahigh vacuum evaporator system at room temperature. For each growth experiment, a catalystcovered substrate was introduced into the HVPE reactor to be heated from room temperature to growth temperature, namely 715 °C. During this 20 min step, the substrates were kept under an undersaturated As2 atmosphere to favor the formation of Ga liquid droplets on the surface and the further wetting of Au with Ga. The GaAs growth was carried out under a H2 vector flow of 3000 SCCM, a III/V ratio (ratio of the GaCl partial pressure to the As2 partial pressure above the substrate) of 5. Note that HVPE experiments can be implemented under III/V ratio either greater than or lower than 1.29 The adsorption/desorption flux ratio of As2g/As4g on GaAs is greater than 1 for partial pressures between 10-4 and 10-3 atm up to 750 °C.28 The surface is thus easily saturated with As elements. This is not the case for chloride molecules that are more volatile. Therefore, a higher pressure of GaCl is required. For information, these experimental parameters yield a GaAs layer-by-layer growth rate of 38 µm/h on vicinal (001) GaAs substrates. The growth time was 15 min. Figure 1 shows the ex-situ SEM images of Au/(001) GaAs substrates after HVPE growth. Perturbated substrate surfaces with a high density of oriented scales are observed from which parallel cylinder-shaped wires continued to grow. Cross-section SEM measurements before and after growth were used to determine the thickness of the rough underlying layer. It has been found to be some 9 µm, corresponding to a growth rate of 36 µm/h. Upon heating prior to growth, one expects decohesion of the Au film, random Au diffusion, and self-assembly into Au particles of various diameters on the substrate surface.11,34-37 This results in a rough Au-Ga surface layer with a significant liquid Ga coverage due to intentional dearth of As on the pregrowth annealing. During 1837

DOI: 10.1021/nl100557d | Nano Lett. 2010, 10, 1836-–1841

ing.30 Zhang et al. observed bending of ceramic nanowires under illumination of electron beam when the current density was high enough.38 Bending can be due to charge accumulation on insulating NWs. HVPE growth of GaAs is carried out with pure Ga (HCl/Ga) and As (AsH3) precursors. The background impurity level of a HVPE reactor is weak (note that the reactor is cleaned under HCl flow at high temperature between each growth experiment). The intrinsic nature of GaAs grown by HVPE is semi-insulating. Therefore the hypothesis of an electrostatic force caused by charge accumulation which bends the NWs is plausible. The wire density and height distribution were measured after SEM on five 2 cm2 samples. The wire density has been estimated at 106 cm-2. The length distribution of the wires is quite narrow. The mean length was measured at 42 µm, including the corrections to account for sample tilt during SEM imaging. The growth rate of the wires is then some 170 µm/h, which is too rapid for condensation. The wires certainly grew by solidification involving a liquid-solid interface fed through the vapor, that is, VLS growth. The mean diameter of the NWs is 120 nm. The diameter of the NWs corresponds to the diameter of the Au particle on top of the NWs (see TEM investigations below). The wires grew parallel to each other at 35° from the substrate (001) plan in the 〈111〉 direction. The 〈111〉B direction should be the energetically favorable one indeed. Finally, let us emphasize that we observed some wires exhibiting unusual length (Figure 1c). Transmission electron microscopy (TEM), electron diffraction, and energy dispersive spectroscopy (EDS) were performed to determine the structure, crystalline orientation, and chemical composition of the wires. GaAs samples were put in ethanol and sonicated for 10 min to separate the wires from the substrates. A few drops of the solutions were then put on a copper grid covered by a holey carbon film. TEM studies were made on a Tecnai G2 ST (Cs ) 1.2 nm), operated a 200 kV, with a point-to-point resolution of 0.25 nm. The observed nanowires have different lengths. This is due to the preparation method for the electron microscopy observations, where nanowires could be broken. Some of the nanowires of GaAs indeed have exceptional lengths. Figure 2 shows a rodlike nanowire roughly 50 µm

FIGURE 1. SEM images at various magnifications (a,b) of rodlike GaAs nanowires grown on (001) GaAs substrates by Au-assisted VLS-HVPE at 715 °C. The growth time was 15 min. The mean length of the wires is 42 µm, the mean diameter is 120 nm and their density is 106 cm-2. The wires are aligned along the 〈111〉 direction (samples were tilted for SEM imaging). Some nanowires were bent during the SEM observation when exposed to the electron beam. (c) SEM image showing a remarkably long wire (130 µm).

HVPE growth, mesoscopic scales corresponding to GaAs nucleation and growth from a Ga rich surface took place with Au-Ga droplets randomly distributed on their faces from which wires were grown. Let us emphasize that nothing has been done to prevent planar growth. As a matter of fact, HVPE does present interest with respect to MOVPE if we can preserve a growth regime governed by rapid surface kinetics. That means working in the 710-750 °C temperature range for GaAs28 and thus giving up the idea of a lowtemperature process aimed at reducing planar growth. In a preliminary conclusion, VLS did take place with success in the high-temperature HVPE environment, and a significant anisotropy was observed between planar and nanowire growth. Note that some wires appear to be bent (Figure 1a). They were bent when the electron beam was focusing on the surface during the SEM observations. Joyce et al. showed as well SEM images of 5 µm long GaAs NWs presenting bend-

FIGURE 2. TEM image of one single nanowire 50 µm long, with constant diameter of 120 nm. © 2010 American Chemical Society

1838

DOI: 10.1021/nl100557d | Nano Lett. 2010, 10, 1836-–1841

FIGURE 5. TEM images of nanowire tips, showing Au particles with different shapes.

FIGURE 3. Electron diffraction patterns obtained on one single nanowire. (a) [1¯10] zone axis in the cubic zinc blende structure. (b) [2¯11] zone axis in the cubic zinc blende structure.

FIGURE 6. Stacking faults in the neck region. The wire was oriented along the [1¯10] direction.

FIGURE 4. HREM image of the edge of one nanowire, taken along the [12¯1] zone axis. An arrow indicates the 〈111〉 growth direction. The distance of the first extinction contours, parallel to the growth direction, allows us to estimate the wedge angles around 60°. Amorphous native oxide covers the nanowire.

A particle of somewhat variable shape is located at the nanowire tip (Figure 5). EDS analyses showed that this particle is nearly pure gold. The growing zone builds a neck; this phenomenon has already been observed by Bauer et al.40 In this region, some stacking faults could be imaged (Figure 6), by orienting the nanowire along a [011] zone axis.41 Apart from this very local part, no stacking faults were detected along the nanowire over the observed 20 µm length, which is a promising result for the crystallographic quality of nanowires grown by HVPE. It is generally assumed that a high growth rate yields lower crystalline quality. Early HVPE experiments demonstrated that high-quality epilayers had been grown at growth rates higher than 100 µm/h.42 Joyce et al. observed that increasing MOVPE precursor flows and consequently the growth rate decreased the NW twin density.30 The outstanding growth rate of Si NWs grown from disilane (31 µm/ min)32 and our observations for the binary GaAs compound suggest that low defect density semiconductor nanowires can be grown very fast. This calls for discussion about the growth mechanisms involved in “rapid VLS” carried out in a VPE environment. VLS growth involves (i) catalytic adsorption and impingement of the growth species (gaseous reactants) at the surface of liquid Au, followed by diffusion of the species to the catalyst-wire interface either through the droplet or on its surface, and crystallization at the liquid-solid interface; (ii) adsorption of the growth species on the substrate surface and diffusion of the ad-species toward the base of the wires, followed by their diffusion along the wire sidewalls until incorporation to the solid-liquid interface.11

long and 120-130 nm wide. The remarkable feature of these nanowires is a constant diameter over outstanding lengths. The nanowires were lying on the carbon-coated grid. Thus their growth direction, found to be 〈111〉, was in plane, and only the zone axis perpendicular to the growth direction could be observed. They were all indexed in the cubic sphalerite (zinc blende) crystallographic structure, which is the normal bulk structure for GaAs. The cubic structure is also that observed for GaAs nanowires presenting diameters greater than 60 nm whatever the growth technique. The most frequent electron diffraction patterns are presented in Figure 3. Figure 3a,b was obtained on the same region of one nanowire by tilting over the 〈111〉 growth direction. Verheijen et al. showed in detail that a common morphology for GaAs nanowires with a 〈111〉 growth direction was hexagonal, with {121} facets.39 Our HREM images are consistent with this observation. Figure 4 shows an HREM image of part of a nanowire, taken along the [1-21] zone axis. The edges of the nanowire correspond to well-defined crystallographic planes. Simulated HREM images matched with Figure 4 for a defocus value of -60 nm (near Scherzer) and for thicknesses varying from 5 to 30 nm. Extinction contours are parallel to the edge of the nanowire and situated at roughly 8 -9 nm from it. These values correspond to wedge angles around 60° and account for a hexagonal faceted nanowire. © 2010 American Chemical Society

1839

DOI: 10.1021/nl100557d | Nano Lett. 2010, 10, 1836-–1841

A VSS mechanism (vapor-solid-solid) was recently discussed for the VPE growth of InAs nanowires involving metal-organic precursors at growth temperatures below the Au-In eutectic melting point.43 The catalyst phase state and composition during growth is not identified in the high temperature AP-HVPE environment. The Au-Ga equilibrium phase diagram shows that for temperature above 700 °C, Au-Ga should be entirely liquid for a high proportion of Ga (20%).44 For process temperature standing below 630 °C, that is, for InP, InAs, and GaAs by MOVPE and MBE, it is established that the group III element determines the catalyst particle state because of the low solubility of the group V element in Au.14 For temperatures above 700 °C, we cannot exclude the introduction of As into the Au-Ga particle during growth, so that the usual binary Au-Ga phase diagram is no longer valid. Taking into account the thermal history and the modification of the equilibrium phase diagram for nanometric objects with respect to bulk material (size-dependent melting temperature depression as predicted45 or in excess of Gibbs-Thompson effect46), we cannot conclude about the catalyst particle without in situ analysis during HVPE growth. Nervertheless, the liquid phase seems plausible (VSS should be excluded) with respect to the process temperature and the observed NW high growth rate; a rich Ga content Au-Ga liquid phase could take place upon the pregrowth annealing of the Au-GaAs substrate, and then the wetting of the catalyst particle with either Ga or both Ga and As could be favored during the NW extension. The postgrowth EDS analysis showed a Ga content below 1% in the tip of grown NWs. Such an after-growth low Ga content can be explained by a quick diffusion of the Ga element to the liquid-solid interface, as a consequence of the concentration gradient established between the droplet surface and the heterogeneous liquid-solid interface where a rapid solidification kinetics yields an almost complete and instantaneous consumption of the Ga species. Rapid Ga incorporation into the droplet is favored by the high dechlorination frequency; GaCl molecules are quickly decomposed, probably as soon as they adsorb on the surface of the Au-Ga droplet. Such a reactant decomposition kinetics-limited growth regime was also demonstrated for the VLS synthesis of millimeter-long Si NWs.32 The observed high GaAs wire growth rate (170 µm/h) indicates that there is no mass hindrance; partial pressures of GaCl and As2 are sufficiently high, and there is a continuous provision to the growing wire. This continuous supply through the catalyst droplet to the liquid-solid interface is responsible for the axial growth of the nanowires. What about radial growth? Radial growth results from direct adsorption of the precursors on the nanowire sidewalls, yielding a lateral thin layer growth. Radial growth can also be fed through material diffusion from the substrate surface toward the wires, followed by a diffusion along the wire sidewalls. The final morphology of the nanowires depends on the growth rate anisotropy between axial and radial growth. In particular, nanowire © 2010 American Chemical Society

tapering is associated to radial growth where a greater fraction of precursor materials reaches the base of the nanowires.47 This diffusion-limited growth is thermally activated. It is obviously observed in MOVPE: the more the tapering, the higher the growth temperature.48 Joyce et al. have synthesized less tapered NWs by promoting high precursor flow rates in order to enhance the axial development of the NWs.30 Unfortunately, the slow decomposition of alkyl-Ga precursors decreases the anisotropy between radial and axial growth. The hypothesis of VLS growth hindered by the low-cracking efficiency of TMGa at low temperature was thoroughly discussed.48,49 At higher temperature, the NW axial growth increases but competes with the thermally activated NW radial growth. In contrast, VLS-HVPE GaAs nanowires have uniform diameter from the base to the top, over length greater than 10 µm. A simple argument excludes any diffusion of matter from the substrate surface along the nanowire sidewalls; the length of the wires is much longer than the surface diffusion length whatever the ad-species. If matter diffused from the substrate surface, then one would have observed a larger base on each of the nanowires. The absence of matter diffusion from the substrate can be explained by the rapid condensation rate on the planar substrate (36 µm/h), which is high enough to incorporate the incoming matter before migration toward the wires. If radial growth takes place in VLS-HVPE, it is fed through uniform adsorption of the precursors directly on the sidewalls of the nanowires. Radial growth involves condensation when axial growth is based on solidification. The growth rate anisotropy between axial and radial growth is high enough to promote preferential axial development of the nanowires and limited radial growth. Let us now return to the link between the growth regime and the crystal homogeneity of the NWs. Changes in the shape of wires were observed when the growth mechanism changed from surface diffusion to direct impingement.14,50 Plante et al. have shown that modifications of the crystal structure of NWs took place where growth by surface diffusion was dominant.11 As already discussed, one expects VLS-HVPE to rely on a preferential axial growth mode through the catalyst droplet. The absence of competition between radial and axial growth, the latter one being very limited, almost nonexistent, favors the crystalline homogeneity of the nanowires and probably the absence of twins. Note that the nanowires synthesized in this study exhibit diameters on the order of 100 nm. Of course, we will not conclude definitely about VLS-HVPE until the synthesis of nanowires with diameters less than 50 nm is achieved. In particular, structure variations, from cubic to wurtzite, were observed for very thin GaAs nanowires grown by MOVPE and MBE. It will be interesting to address the VLS-HVPE growth of thin NWs (on controlled prestructured Au) to assess the competition between the two crystallographic structures if preferential axial growth mode is preserved. 1840

DOI: 10.1021/nl100557d | Nano Lett. 2010, 10, 1836-–1841

(17) Kuykendall, T.; Pauzauskie, P. J.; Zhang, Y.; Goldberger, J.; Sirbuly, D.; Denlinger, J.; Yang, P. Nat. Mater. 2004, 3, 524. (18) Joyce, H. J.; Gao, Q.; An, H. H.; Jagadish, C.; Kim, Y.; Zhang, X.; Guo, Y.; Zou, J. Nano Lett. 2007, 7, 921. (19) Skold, N.; Wagner, J. B.; Karlsson, G.; Hernan, T.; Seifert, W.; Pistol, M. E.; Samuelson, L. Nano Lett. 2007, 12, 2743. (20) Johansson, J.; Karlsson, L. S.; Svensson, C. P. T.; Martensson, T.; Wacaser, B. A.; Deppert, K.; Samuelsson, L.; Seifert, W. Nat. Mater. 2006, 5, 574. (21) Paiano, P.; Prete, P.; Speiser, E.; Lovergine, N.; Richter, W.; Tapfer, L.; Mancini, A. M. J. Cryst. Growth 2007, 298, 620. (22) Bhunia, S.; Kawamura, T.; Fujikawa, S.; Nakashima, H.; Furukawa, K.; Torimitsu, K.; Watanabe, Y. Thin Solid Films 2004, 464-465, 244. (23) Ohlsson, B. J.; Bjo¨rk, M. T.; Magnusson, M. H.; Deppert, K.; Samuelson, L.; Wallenberg, L. R. Appl. Phys. Lett. 2001, 3335, 79. (24) Samuelson, L.; Thelander, C.; Bjo¨rk, M. T.; Borgstro¨m, M.; Deppert, K.; Dick, K. A.; Hansen, A. E.; Martensson, T.; Panev, N.; Persson, A. I.; Seifert, W.; Skold, N.; Larsson, M. W.; Wallenberg, L. R. Physica E 2004, 25, 313. (25) Martensson, T.; Calberg, P.; Borgstro¨m, M.; Montelius, L.; Seifert, W.; Samuelson, L. Nano Lett. 2004, 4, 699. (26) Lassalle, F.; Porte, A.; Laporte, J. L.; Pariset, C.; Cadoret, R. Mater. Res. Bull. 1988, 23, 1285. (27) Pimpinelli, A.; Cadoret, R.; Gil-Lafon, E.; Napierala, J.; Trassoudaine, A. J. Cryst. Growth 2003, 258, 1. (28) Gil-Lafon, E.; Napierala, J.; Pimpinelli, A.; Cadoret, R.; Trassoudaine, A. J. Cryst. Growth 2003, 258, 14. (29) Gil-Lafon, E.; Napierala, J.; Castelluci, D.; Pimpinelli, A.; Cadoret, R.; Gerard, B. J. Cryst. Growth 2001, 222, 482. (30) Joyce, H. J.; Gao, Q.; Hoe Tan, H.; Jagadish, C.; Kim, Y.; Fickenscher, M. A.; Perera, S.; Hoang, T. B.; Smith, L. M.; Jackson, H. E.; Yarrison-Rice, J. M.; Zhang, X.; Zou, J. Nano Lett. 2009, 9, 695. (31) Morales, A. M.; Lieber, C. M. Science 1998, 279, 208. (32) Park, W.; Zheng, G.; Jiang, X.; Tian, B.; Lieber, C. M. Nano Lett. 2008, 8, 3004. (33) Cui, Y.; Wei, Q.; Park, H.; Lieber, C. M. Science 2001, 293, 1289. (34) Plante, M. C.; Garrett, J.; Ghosh, S. C.; Kruse, P.; Schriemer, H.; Hall, T.; LaPierre, R. R. Appl. Surf. Sci. 2006, 253, 2348. (35) Dubrovskii, V. G.; Cirlin, G. E.; Soshnikov, I. P.; Tonkikh, A. A.; Sibirev, N. V.; Samsonenko, Y. B.; Ustinov, V. M. Phys. Rev. B. 2005, 71, 205325. (36) Bauer, C. L. Surf. Sci. 1986, 168, 395. (37) Holloway, P. H.; Mueller, C. H. Thin Solid Films 1992, 221, 254. (38) Zhang, Y.; Zhu, J. Micron 2002, 33, 523. (39) Verheijen, M. A.; Algra, R. E.; Borgstro¨m, M. T.; Immink, G.; Sourty, E.; Van Eckervort, J. P.; Vlieg, E.; Bacckers, E. P. A. M. Nano Lett. 2007, 7, 3051. (40) Bauer, J.; Gottschalch, V.; Paetzelt, H.; Wagner, G.; Fuhrman, B.; Leipner, H. S. J. Cryst. Growth 2007, 298, 625. (41) Karlsson, L. S.; Dick, K. A.; Wagner, J. B.; Malm, J. O.; Deppert, K.; Samuelson, L.; Wallember, L. R. Nanotechnology 2007, 18, 48717. (42) Lourdudoss, S.; Gopalakrishnan, N.; Holz, R.; Deschler, M.; Beccard, R. Metall. Mater. Trans. A 1999, 30, 1047. (43) Dick, K. A.; Deppert, K.; Martensson, T.; Mandl, B.; Samuelson, L.; Seifert, W. Nano Lett. 2005, 5, 761. (44) Massalski, T. B. Binary alloy phase diagrams; American Society for Metals: 1990; Vol. 1. (45) Kodambaka, S.; Tersoff, J.; Reuter, M. C.; Ross, F. M. Science 2007, 316, 729. (46) Park, H. D.; Gaillot, A. C.; Prokes, S. M.; Cammarata, R. C. J. Cryst. Growth 2006, 296, 159. (47) Kim, Y.; Joyce, H. J.; Gao, Q.; Tan, H. H.; Jagadish, C.; Paladugu, M.; Zou, J.; Suvorova, A. A. Nano Lett. 2006, 6, 599. (48) Borgstro¨m, M.; Deppert, K.; Samuelson, L.; Seifert, W. J. Cryst. Growth 2004, 260, 18. (49) Fortuna, S. A.; Wen, J.; Chun, I. S.; Li, X. Nano Lett. 2008, 8, 4421. (50) Dubrovskii, V. G.; Soshnikov, I. P.; Cirlin, G. E.; Tonkikh, A. A.; Samsonenko, Y. B.; Sibirev, N. V.; Ustinov, V. M. Phys. Status Solidi B 2004, 241, R30.

To summarize, the feasibility of Au-assisted VLS growth of GaAs nanowires has been shown in a HVPE environment. Forty micrometer long NWs were synthesized in 15 min on (001) substrates at a density of 106 cm-2 from a uniform Au-Ga surface layer simply structured upon the 20 min pregrowth annealing in the HVPE reactor. The direction of the wires was 〈111〉B and their structure was cubic zinc blende. The mean growth rate of the nanowires was 170 µm/ h. Such an order of magnitude emphasizes that the growth mechanism of GaAs NWs by HVPE is mainly based on solidification after direct and continuous feeding of the growth species from the vapor phase through the Au-Ga liquid alloy catalyst. The high solidification rate is facilitated by the high-decomposition frequency of the chloride growth precursors involved in the HVPE process. The nanowires exhibited rodlike morphology. Their constant diameter, as well as their length being greater than 10 µm, supports the assumption of direct feeding through the liquid-solid interface and excludes ad-species diffusion-limited growth, either from the substrate surface or the sidewalls of the NWs. The availability of long III-V NWs opens up new opportunities. Applications that were already addressed for Si, such as enhanced integration through the definition of a large number of devices on a single NW, multifunctionalized biosensors, or multiplexed biodetectors, can now be investigated for the III-V. Acknowledgment. We thank Professor Philip Hoggan for his careful reading of the text and language corrections. REFERENCES AND NOTES (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)

Duan, X. F.; Huang, Y.; Agarwal, R.; Lieber, C. M. Nature 2003, 421, 241. Park, H. G.; Barrelet, C. J.; Wu, Y.; Tian, B.; Qian, F.; Lieber, C. M. Nat. Photon. 2008, 2, 622. Qian, F.; Li, Y.; Gradee`ak, S.; Barrelet, C. J.; Wang, D.; Lieber, C. M. Nano Lett. 2004, 4, 1975. Hahm, J. I.; Lieber, C. M. Nano Lett. 2004, 4, 51. Ng, H. T.; Han, J.; Yamada, T.; Nguyen, P.; Chen, Y. P.; Meyyappan, M. Nano Lett. 2004, 4, 1247. Sun, Y.; Khang, Y.; Hua, F.; Hurley, K.; Nuzzo, R. G.; Rogers, J. A. Adv. Func. Mater. 2005, 15, 30. Motohisa, J.; Noborisaka, J.; Takeda, J.; Inari, M.; Fukui, T. J. Cryst. Growth 2004, 272, 180. Tomioka, K.; Motohisa, J.; Hara, S.; Fukui, T. Nano Lett. 2008, 8, 3475. Wagner, R. S.; Ellis, W. C. Appl. Phys. Lett. 1964, 4, 89. Hiruma, K.; Katsuyama, T.; Ogawa, K.; Koguchi, M.; Kakibayashi, H.; Morgan, G. P. Appl. Phys. Lett. 1991, 59, 431. Plante, M. C.; Lapierre, R. R. J. Cryst. Growth 2006, 286, 394. Piccin, M.; Bais, G.; Grillo, V.; Jabeen, F.; De Franceschia, S.; Carlinoa, E.; Lazzarino, M.; Romanato, F.; Businaro, L.; Rubini, S.; Martelli, F.; Franciosi, A. Physica E 2007, 37, 134. Ihn, S. G.; Song, J. I. Nanotechnology 2007, 18, 355603. Tchernycheva, M.; Travers, L.; Patriarche, G.; Glas, F.; Harmand, J. C.; Cirlin, G. E.; Dubrovskii, V. G. J. Appl. Phys. 2007, 102, No. 094313. Jensen, L. E.; Bjork, M. T.; Jeppesen, S.; Persson, A. I.; Ohlsson, B. J.; Samuelson, L. Nano Lett. 2004, 4, 1961. Sivakov, V. V.; Heyroth, F. F.; Falk, F. F.; Andr, G. G.; Christiansen, S. S. J. Cryst. Growth 2004, 272, 167.

© 2010 American Chemical Society

1841

DOI: 10.1021/nl100557d | Nano Lett. 2010, 10, 1836-–1841