Home
Search
Collections
Journals
About
Contact us
My IOPscience
Catalytic growth of ZnTe nanowires by molecular beam epitaxy: structural studies
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2007 Nanotechnology 18 475606 (http://iopscience.iop.org/0957-4484/18/47/475606) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 202.43.188.4 This content was downloaded on 30/09/2013 at 02:57
Please note that terms and conditions apply.
IOP PUBLISHING
NANOTECHNOLOGY
Nanotechnology 18 (2007) 475606 (8pp)
doi:10.1088/0957-4484/18/47/475606
Catalytic growth of ZnTe nanowires by molecular beam epitaxy: structural studies E Janik1, P Dłu˙zewski1 , S Kret1 , A Presz2 , H Kirmse3 , W Neumann3 , W Zaleszczyk1 , L T Baczewski1 , A Petroutchik1 , E Dynowska1 , J Sadowski1 , W Caliebe4 , G Karczewski1 and T Wojtowicz1 1
Institute of Physics, Polish Academy of Sciences, aleja Lotnik´ow 32/46, 02 660 Warszawa, Poland 2 Institute of High Pressure Physics (UNIPRESS), Polish Academy of Sciences, ulica Sokołowska 29/37, 01-142 Warszawa, Poland 3 Institut f¨ur Physik, Humboldt-Universit¨at zu Berlin, AG Kristallographie, Newtonstrasse 15, D-12489 Berlin, Germany 4 Hasylab at DESY, Notkestrasse 85, D-22603 Hamburg, Germany E-mail:
[email protected]
Received 2 August 2007, in final form 4 October 2007 Published 26 October 2007 Online at stacks.iop.org/Nano/18/475606 Abstract ZnTe nanowires were grown by molecular beam epitaxy on GaAs substrates of three different orientations: (100), (110), and (111)B. The catalyst droplets were produced through in situ annealing of a previously deposited Au layer and by forming the eutectic alloy with Ga from the substrate. The influence of substrate orientation and growth parameters on the properties of nanowires was investigated using scanning and transmission electron microscopy, energy dispersive x-ray spectroscopy, and x-ray diffraction. The growth process was based on the vapour–liquid–solid mechanism and the contribution of the diffusion-induced effect in this mechanism was confirmed by correlating the length and the diameter of the produced nanowires. The nanowires had diameters ranging from 30 to 70 nm and lengths between 1 and 2 μm. The growth axis of the nanowires was 111 and the nanowires grew along 111 directions of the substrate, independent of the substrate orientation used. The nanowires had stacking faults at the bottom and those grown at optimal conditions possessed perfect cubic structure near the top. (Some figures in this article are in colour only in the electronic version)
working as biochemical sensors [8], light-emitting diodes [9] and lasers [2, 10], resonant tunnelling diodes [11], field effect transistors [5], single-electron transistors [12], and photodetectors [13] have been demonstrated recently. Modern crystal growth technologies such as molecular beam epitaxy (MBE) and metallorganic vapour phase epitaxy (MOVPE) provide a feasible way for material engineering by implementing various material combinations together with well-controlled doping levels in individual nanowires. Thus, diverse junctions can be formed either by sequential deposition of n- and p-type doped materials during the growth or by crossing methods [5]. A great selection of semiconducting materials has been successfully tested for NW growth,
1. Introduction One-dimensional (1D) nanostructures have become the focus of many research laboratories [1–4] over the last years, not only because of their interesting electronic and optical properties associated with lower dimensionality that result in pronounced quantum confinement effects, but also because of their potential for applications [5]. The prospect of using semiconductor nanowires (NWs) as ‘building blocks’ for nanoscale electronic and photonic devices [6, 7] has sparked intensive research efforts aimed at understanding the formation processes as well as describing the structural properties of this class of nanostructures. NW-based devices 0957-4484/07/475606+08$30.00
1
© 2007 IOP Publishing Ltd Printed in the UK
Nanotechnology 18 (2007) 475606
E Janik et al
including Si, Ge [14–17], III–V compounds [18–21] and II– VI compounds [22–29]. The idea of growth of nanowires is inspired by the relatively old mechanism of catalytic growth of silicon whiskers of micrometre dimensions developed by Wagner and Ellis in 1965 [30]. This method is known as the vapour– liquid–solid (VLS) mechanism and includes using a catalyst in the shape of droplets. Exact growth mechanisms involved in catalytic growth of 1D structures with nanometric dimensions are not yet completely understood and are still debated. Particular attention has been paid to analysing the dependence of the NW growth rate on their diameter. A theoretical model was recently developed by Schubert et al [14] and Dubrovskii et al [31, 32] for MBE growth, and by Johansson et al [33] for MOVPE growth. There are many reports of NWs grown from selenides, sulfides and, recently, oxides (ZnO). As for tellurides, CdTe NWs [25, 26] and ZnTe NWs [27–29] have been obtained by chemical synthesis or CVD methods. Developing reliable methods of growing nanowires made of II–VI semiconductors, particularly ZnTe, is quite important for several reasons. ZnTe is a direct band gap ( E g = 2.4 eV) semiconductor and it possesses properties that are attractive for both basic research and potential applications. In particular, ZnTe can be very heavily p-type doped with nitrogen (N), which allows obtaining the highest free hole concentration among all other II–VI materials. In addition, II–VI tellurides are the best known diluted magnetic semiconductors (DMSs) [34, 35]. Therefore they can provide an appealing bridge between ‘spintronics’ [36] and the ‘bottom up’ approach for nanostructure formation [5]. Finally, NWs made of tellurides are expected to become a basis for obtaining magnetic and p-type doped oxide NWs via the oxidation route (e.g. p-ZnMnO from ZnMnTe:N). Our preliminary results on ZnTe NWs grown on GaAs (100) were presented recently [37]. In this paper we describe in detail the structural properties of ZnTe nanowires grown by MBE on the GaAs substrates of three different orientations and the analysis of the characteristics of the underlying growth mechanism. We used (100), (110) and (111)B oriented substrates to show that the method of making catalyst nanoparticles by prolonged, high temperature heating of the Au layer inevitably leads to the growth of ZnTe NWs having 111 growth axis, which is, most often, the preferred growth axis of nanowires made of other materials as well [38–40], although the MBE growth of 110 and 112 oriented ZnSe NWs was reported for sizes of Au nanoparticles smaller than 20 nm [41, 42]. We carefully investigated the orientation of nanowires relative to the orientation of the substrate to prove that the 111 growth axis of NWs is oriented along 111 directions of the substrate, independently of the substrate orientation used, and hence that there is an epitaxial relation between the NWs and the substrate. Based on EDX analysis of the chemical composition of the nanosphere located at the tip of each nanowire, we confirmed a previously proposed explanation why the epitaxial growth in the direction other that the substrate normal is possible [43]. Finally, by performing an analysis of the relation between the NW length and diameter, we provide strong evidence that the diffusion-induced mechanism, proposed by Dubrovskii et al [31, 32] and Johansson et al [33], plays a crucial role in the MBE growth process of ZnTe NWs.
2. Experimental details The growth of ZnTe nanowires was performed in an EPI 620 MBE system, using a GaAs substrate with a gold layer deposited in a separate MBE chamber. Prior to the growth of NWs the nanocatalysts were formed by in situ annealing of the substrate. Three types of substrates used for the current studies, (100), (110) and (111)B oriented GaAs/Au, were indium mounted on the same molybdenum block. This assures the same growth conditions and allows for direct comparison of NWs grown on substrates with different orientation. As an initial stage of our research the influence of the annealing temperature on the formation of Au-based nanoparticles on GaAs substrate was studied. The MBE system was equipped with low temperature effusion cells and a reflection high energy electron diffraction (RHEED) system. The growth was performed from elemental Zn and Te sources. The growth temperature was controlled by a thermocouple and the real substrate temperature was obtained from the calibration based on the melting points of In, Sn, and Pb. The following MBE growth parameters were thoroughly investigated during NW growth: the substrate temperature Tg , the density of impinging fluxes, and the relative flux ratios for a given growth time. For the characterization of the nanowires, field emission scanning electron microscopy (FE-SEM: Leo 520), high resolution transmission electron microscopy (HRTEM: JEOL 2000EX), energy-dispersive x-ray spectroscopy (EDXS: R¨ontec system attached to a JEOL 2200FS), and xray diffraction (XRD) were used. X-ray measurements were performed in symmetrical ω–2θ scan mode with the ˚ produced monochromatic beam of wavelength λ = 1.540 56 A at the synchrotron W1.1 beam line at DESY-HASYLAB. FE-SEM pictures were taken on pieces of ‘as grown’ samples, cleaved off the original substrates. Most of the TEM studies were performed on specimens gained by scraping the surface of the sample with a TEM grid, coated with perforated carbon film, in order to separate the NWs from the substrate. For the cross-sectional observations, the sample was embedded in epoxy glue and prepared using mechanical polishing and dimpling, followed by Ar ion milling.
3. Results and discussion In our studies of the formation of nanocatalysts we have found that the diameter of Au-based droplets after heat treatment of the Au layer, with the thickness in the range from 0.3 to 2 nm, depends predominantly on the annealing time, and only weakly on the thickness of the initial layer. We have not observed any significant dependence of the average droplet diameter on substrate orientation. The diameters d of the nanoparticles observed after cooling down the substrate to room temperature were approximately 20 nm for 5 min, 50 nm for 10 min and 100 nm for 15 min annealing time (figures 1(a)–(c)). Our final choice for the NW growth reported in this paper was 1 nm of Au layer thickness and 10 min of annealing time at 550 ◦ C. The resulting density of Au-based droplets was of the order of 1 × 109 cm−2 . As we will discuss later, on the basis of the results of our EDX studies, nanodroplets are formed by making Au–Ga eutectic, with Ga originated from the GaAs substrate. In the process of prolonged annealing additionally 2
Nanotechnology 18 (2007) 475606
E Janik et al
Figure 1. FE-SEM images of Au–Ga nanoparticles on GaAs substrate resulting from the annealing of a 1 nm thick Au layer for various times at 550 ◦ C: (a) d ∼ 20 nm for 5 min, (b) d ∼ 50 nm for 10 min, (c) d ∼ 100 nm for 15 min, where d is the diameter of the nanoparticle.
Figure 2. FE-SEM images of ZnTe NWs grown on GaAs substrates with various orientations. Left column—top view images for (111)B, (110), and (100) oriented substrates: (a), (b) and (c), respectively. Right column—side view images for (111)B, (110), and (100) oriented substrates: (d), (e) and (f), respectively. For each panel in the right column the nominal viewing directions is given. Samples presented in (a)–(c) were grown simultaneously in the same growth process. Also samples from (d) and (e) were grown simultaneously in another growth process. The insets show schematic projections of NWs grown along 111-type crystallographic directions of the substrate onto the plane of the pictures.
the As evaporates and low energetic {111}-type surfaces are revealed, leading also to the strong corrugation of the GaAs surface. This surface corrugation is especially well visible
in figure 1(c) for the (110) oriented substrate, since for this orientation only two {111}-type surfaces exist and therefore elongated trenches are formed. 3
Nanotechnology 18 (2007) 475606
E Janik et al
Figure 3. The x-ray diffraction pattern obtained for ZnTe nanowires grown on (100) oriented GaAs substrate in a symmetrical ω–2θ scan.
Figure 5. Length of nanowires grown on two different GaAs substrates oriented (110) and (111)B plotted versus the nanowire diameter as measured in the middle of its length. The inset shows the length of the nanowires as a function of Zn and Te fluxes for constant Zn/Te ratio equal to about 0.6. Lines are the least square fits of the L = A/d relation, separated for (110) and (111)B growth, where L is the NW length, A is a fit constant, and d is the NW diameter.
type directions of the substrate would look like in the picture. In these insets NWs aligned along particular directions are represented by the white rods with the balls at their tips and are labelled accordingly. Figures 2(a) and (d) reveal that the majority of nanowires grown on (111)B oriented substrate are vertical and some limited number form an angle with the substrate (see figure 2(d)). Those NWs which grow vertically—in the [111]B direction of the substrate—are visible in figure 2(a) as white dots (projections of the NWs) with black centres (nanocatalysts). The projections of nanowires that are not vertical to the surface are seen in figure 2(a) to be aligned along three directions forming a 120◦ angle among each other, as expected for NWs that grew along three 111A-type directions of the substrate. Both 111B-type and 111A-type NWs are visible also in the side view of the sample presented in panel (d), and they can be identified by comparing with the schematic picture in the inset. The 111A-type nanowires form an angle of 19.5◦ with the substrate, but their projections onto the picture plane make either a 19.5◦ or 35.3◦ angle with the substrate edge. Figure 2(b) shows that for the growth on the (110) surface the projections of practically all NWs are aligned along parallel lines and that the NWs have a tendency to point in one direction (to the down-right, as can be judged by looking at the positions of the balls at the NW tips). Also in the panel (e) it is seen that the NWs form a very specific angle with the substrate surface (shown to be 54.7◦ in the careful TEM studies discussed later), and that the number of NWs pointing up-right is larger then those pointing up-left. This indicates that NWs on the (110) substrate grow preferentially along the [111¯ ]B direction and some along the [111]A direction. These directions are the only two directions belonging to the 111 family that stick out from the (110) oriented GaAs substrate. For the growth of NWs on (100) GaAs substrate (see figures 2(c) and (f)), as already reported in [37], four preferred
Figure 4. The distribution of nanowires by diameter of Au-based particles at the tip of NWs grown on (111)B GaAs substrate.
Having established the procedure for Au–Ga droplet formation, we investigated the influence of the principal parameters of the growth process on the formation of NWs. The growth parameters were optimized so as to grown nanowires that were reproducible from growth to growth, homogenous in shape, straight, and contained a minimum number of stacking faults. We have found that the optimal parameters of growth of nanowires were as follows: Tg = (360–450 ◦ C), Zn/Te beam equivalent pressure (BEP) about 0.65. Figures 2(a)–(c) show the top view FESEM images of the ZnTe NWs grown simultaneously in the same growth process on (111)B, (110) and (100) oriented Au/GaAs substrates. The growth temperature was 430 ◦ C. In figures 2(d)–(f) we show side views of samples grown at similar growth conditions on substrates with analogous orientations. The choice of the samples was made so that the inclination of NWs to the substrate surface is clearly visible in their images. The insets in all panels are schematic drawings showing how the NWs that grew along one of the 1114
Nanotechnology 18 (2007) 475606
E Janik et al
¯ ] direction is perpendicular Figure 6. Cross-sectional TEM image of the sample grown on (110) substrate. The GaAs surface plane [111 to the viewing plane; out of the plane [111]A and [111¯ ]B directions of NWs are projected onto the image plane as [121] and [211¯ ] respectively.
Figure 7. Shape analysis of NWs: (a) visualization of the stacking fault plane geometry, (b) magnification of the bottom part of the nanowire shown in panel (a), (c) bottom part of a nanowire with high density of stacking faults, (d) TEM image of a nanowire taken with the electron beam almost parallel to the axis of the NW, (e) processed TEM image of the NW’s projection along its axis, (f) electron diffraction pattern corresponding to (e), (g) suggested evolution of the shape of the cross-section of the nanowires.
orientations of NWs are observed, since there are four 111type directions available. The top view image (cf figure 2(c)) shows that the NWs are randomly distributed on the (100)GaAs substrate and their projections on the substrate surface are preferentially aligned along two respectively perpendicular directions parallel to the cleavage edges of GaAs: [011] and ¯ ]. The side view (panel (f)) reveals that NWs form an angle [011 of 35.3◦ with the substrate surface, as they should if they grew along 111 directions of GaAs. The conclusion from the FE-SEM studies is that, independently of the orientation of the GaAs substrate, ZnTe NWs grow preferentially along 111-type directions of the substrate, with the number of 111B oriented nanowires dominating over those oriented along 111A directions. The SEM results also strongly suggest that the preferred growth axis of ZnTe NWs must be 111. This 111 growth axis of NWs was proven by two methods: XRD and high resolution TEM. These experiments also provide strong proof of the epitaxial relation between the substrate and the growing NWs.
XRD measurements were realized with a monochromatic ˚ in symmetrical x-ray beam of wavelength λ = 1.540 56 A ω–2θ scans. In this mode of measurement the detector position (2θ angle) is coupled with the maximum intensity of the suitable rocking curve (ω angle) resulting from the crystallographic orientation of the GaAs substrate. Such a measurement allows detecting the lattice planes parallel to the crystallographic orientation of the substrate. The example of the diffraction pattern obtained for nanowires grown on (100)-GaAs substrate is shown in figure 3. Analysis of this pattern shows that the crystallographic orientation of the substrate imposes the orientation of the nanowires: the strongest among observed reflections of ZnTe are indexed 002, 004 and 006, respectively, and the correspond to analogous reflections of the GaAs substrate. This means that the (100) lattice planes in the nanowires are parallel to the (100) lattice planes of the substrate. If so, the nanowires growing along 111 crystallographic directions (TEM results) have to form an angle of 35.3◦ with the (100) lattice planes 5
Nanotechnology 18 (2007) 475606
E Janik et al
of the substrate according to the interplanar angles in cubic crystals. The analysis of nanowires grown on (110) and (111) oriented GaAs leads to similar conclusions: in the first case the nanowires form an angle of 54.7◦ with the (110) lattice plane of the substrate, while the nanowires grown on (111) oriented substrate are either perpendicular to the (111) lattice plane of the substrate or form an angle of 19.5◦ with the (111) lattice plane. Additional small peaks visible in figure 3 (e.g. indexed 111, 220, 311), as confirmed by an additional experiment performed in the grazing incidence geometry, originate mainly in the thin polycrystalline layer of ZnTe that forms directly on the GaAs substrate during the growth of NWs (see also figure 6). The typical dimensions of nanowires determined from TEM observations and FE-SEM images are as follows: 30– 70 nm diameter at the top of wires and 0.7–2.6 μm length for 30 min growth duration (dependent on the flux of atoms— see below). The distribution by diameter of particles at the tip of nanowires grown on GaAs (111)B substrate is presented in figure 4. We have observed that all the wires grown in our experiments were terminated by Au-based particles of quasi-spherical form. This confirms the hypothesis that the well-known VLS mechanism is responsible also for the MBE growth of ZnTe NWs. The tapering of all nanowires is clearly seen in our electron microscopy studies (see figures 2, 6 and 7); however the reason for the wire tapering is not fully understood. In order to better understand the mechanisms of MBE growth of NWs and to verify the contribution to the growth from the diffusion of adatoms on the sidewalls of the nanowires we have determined the correlation between the length L and the diameter d of ZnTe nanowires. We have observed that thinner NWs grow faster than the thicker ones. This is shown in figure 5 for two samples grown on (111)B and (110) oriented substrates in different growth runs. The NWs on (111)B substrate were grown with the Zn flux having BEP of 5 × 10−7 Torr and those on (110) with flux of 7.6 × 10−7 Torr. In both growths, independently of the particular growth rate, a 1/d dependence of the lengths of nanowires on their diameter is observed. This dependence reported previously for a number of NWs made of various materials and grown both by the MOVPE [44], chemical beam epitaxy (CBE) [3], and MBE [14] is considered to be evidence for the strong contribution to the growth of NWs from the atoms impinging not directly onto the eutectic ball but either on the substrate or on the NWs’ side surface. Those atoms then diffuse along the sidewalls of the NW toward the eutectic ball, dissolve in it and contribute to the growth of the NW at the boundary between the supersaturated liquid and the NW. Of course only the adatoms that impinge within a distance from eutectic drop smaller than a surface diffusion length can contribute. However, for the typical, relatively high substrate temperatures used for catalytically enhanced NW growth, this length can reach a couple of micrometres [44]. In our growth experiments we also tested the influence of Zn and Te fluxes on the length of nanowires for a constant growth time tg = 30 min. We find that increasing the Zn and Te fluxes by about 25% without changing their relative Zn/Te ratio resulted in a proportional increase of the nanowire length by 100%, as is shown in the inset of figure 5.
Figure 8. Two types of NW structure near the top. (a) Type A, high density of the stacking faults with thin twin segments, (b) type B, top part of NW free of defects; (c) and (d) magnified views of (a) and (b) respectively.
The theoretical model of the NW formation by the VLS mechanism during molecular beam epitaxy which includes a diffusion-induced contribution was developed by Dubrovskii et al [31, 32]. This theory predicts that the length of NWs growing primarily due to adatoms diffusion is inversely proportional to the diameter of the eutectic drop. On the other hand if the diffusion-induced effects can be neglected then the thicker nanowires should grow faster. Therefore, our results presented in figure 5 provide strong evidence for the important contribution to the growth of NWs from adatoms diffusing on the sidewalls of the ZnTe NWs during the MBE process. It is worth mentioning that Wang and Fishman [45] considered yet another type of surface diffusion process that is important for nanowire growth by the VLS mechanism. Those authors have shown theoretically that diffusion along a liquid alloy droplet surface can constitute a vital mass transport channel, especially in the case of compound NWs, that becomes more important with decreasing nanowire diameter. However, the explicit discussion of the NWs’ growth rate as a function of their diameter was not provided. Detailed structural information on the ZnTe nanowires was obtained from high resolution transmission electron microscopy. Electron diffraction from a large number of the NWs deposited on the perforated carbon film reveals ring patterns characteristic for the sphalerite crystal structure with lattice constant of 0.61 nm. High resolution images and the electron diffraction pattern of single NWs confirm this crystal structure and show that the growth axis of nanowires is always 111, independent of whether the NWs were grown on (111)B, (110) or (100) oriented substrates. The crosssectional image shown in figure 6 provides information on the 6
Nanotechnology 18 (2007) 475606
E Janik et al
grown at optimal conditions have almost perfect structure in the upper part (type B). Figure 8 shows the HRTEM images of these two types of nanowire. As already mentioned in the introduction, the fact that the growth axis of ZnTe NWs is 111 is consistent with the preferred growth axis of NWs made of many other semiconductor materials. However, ZnTe NWs grown epitaxially on (100) and (110) oriented GaAs substrates would be expected to have growth axis 100 and 110, respectively, by analogy with the epitaxy of the layered structure. The epitaxy in 111 directions on (100) and (110) substrates can only be understood in the case when the GaAs substrate is locally dissolved in the reaction with the Au layer [43], and low energy {111} facets develop within the pit. Such facets can be a starting point for the epitaxial growth of 111 oriented NWs, observed in our experiments for all substrate orientations. The strong confirmation of this scenario is the experimental demonstration that the nanoparticles present at the tips of our NWs are a solidified eutectic alloy of Au with Ga, originated from dissolved GaAs. This was done by performing studies of the chemical composition of the semi-spherical catalytic particle that is present at the tip of each nanowire with the use of the EDXS. The EDXS analysis was done for the individual nanowires deposited on the holey carbon film supported by a copper grid. The size of the electron probe used in this experiment was 1 nm. The composition analysis was performed applying Quantax 1.8 software. The results of this semi-quantitative analysis of the catalytic particle were as follows: 75 at.% of Au, about 22 at.% of Ga, and about 3 at.% of Zn. The 4 nm thick shell of the sphere is composed of about 40 at.% of Au, 50 at.% of Ga and also about 7 at.% of Te, and 3–4 at.% of Zn. The EDXS line scans for the same sample were performed on the cross-section geometry (the nanowires were embedded in epoxy glue together with the substrate). In figure 9(a) the TEM image shows the NW after the line scan, with the trace of the electron beam visible as a darker line mainly in the glue area, (cf arrows). The nominal spot size was, again, about 1 nm. The curves in figure 9(b) clearly show an abrupt change in composition at the flat interface between the semi-spherical catalyst and ZnTe NW for all analysed elements. These EDXS results strongly confirm the model of the formation of Au-based eutectic particles by dissolving GaAs and exposing {111} facets for the epitaxial growth of 111 oriented ZnTe NWs [43].
Figure 9. EDXS analysis of the top part of an NW: (a) TEM image of the area used for EDXS analysis, (b) EDXS line scan across catalytic sphere and NW (the intensity of Ga signal was corrected by subtracting the background signal). Within the resolution power of the technique the noise seen for the individual curves has no physical relevance.
substrate/nanowire interface and additionally on the orientation of the NWs’ growth axis in respect to the crystallographic directions of the substrate (the later are more precise than those gained from FE-SEM). Analysis of the angles observed in figure 6 for NWs formed on the (110) substrate proves that NWs grow practically only along two orientations: [111]A and [111¯ ]B out of the substrate plane. Additionally, as was visible also in figure 2(b) for another sample, one can notice (in figure 6) that the number of NWs that grew along the [111¯ ]B direction is larger than the number of [111]A oriented NWs. In figure 6 one can also clearly see that the thickness of the ZnTe layer that was formed on GaAs substrate during the growth of NWs is about 50 nm, and is therefore much smaller than the length of NWs, which exceeds 1 μm. Therefore, around a twenty-fold enhancement of the linear growth rate of NWs in respect to the two dimensional growth was achieved with the use of the Au-based nanocatalyst. The details of the bottom part geometry of the NWs are shown in figure 7. For the majority of nanowires the bottom part reveals the sawtooth-like form of the sidewalls correlated with a high number of stacking faults near the interface between the nanowire and the substrate. Figure 7 shows also that the shape of visible contours of the stacking faults evolves from triangular in the bottom part to hexagonal at the top. Figures 7(c) and (d) give the image of an NW in almost vertical projection. Figure 7(e) presents the corresponding selected area electron diffraction (SAED) pattern typical for the 111 zone axis. The facets are formed by the {110}type planes parallel to the 111 zone axis. The triangular or hexagonal shape can be mutually transformed during the growth by cutting the vertex of the triangles by {110} planes, as is schematically shown in figure 7(f). The tops of the nanowires typically have a lower number of defects than their bottom parts. Some nanowires, grown at non-optimal conditions, contain stacking faults perpendicular to the axis of growth up to the very top (type A) but most NWs
4. Conclusions ZnTe nanowires were grown by MBE on GaAs substrates with three different orientations. The nanowires have 111 growth axis oriented along 111 directions of the GaAs substrate, regardless of the substrate orientation. Epitaxy in the direction other than the substrate normal is possible due to the formation of the {111} facets during the process of annealing of Au. This annealing leads to the formation of Au/Ga eutectic nanodroplets with Ga originated from the substrate. Strong confirmation of this scenario is provided by EDX analysis of the chemical composition of the nanosphere located at the tip of each nanowire revealing the presence of 22% of Ga. The relation between parameters of growth 7
Nanotechnology 18 (2007) 475606
E Janik et al
and the structural properties of the nanowires were analysed. The dependence of NW length on NW diameter confirms the strong contribution from a diffusion-induced mechanism in VLS growth of ZnTe NWs. The nanowires grown at optimal conditions possessed perfect cubic structure near the top and show the formation of stacking faults or microtwins at the bottom. The growth of completely defect-free 100 or 110 oriented NWs would require either the use of other methods of producing a nanocatalyst, not leading to the dissolving of the GaAs substrate, and hence allowing one to force the epitaxial growth along a direction perpendicular to the (100) or (110) oriented substrates, or, possibly, the use of the sizeand/or temperature-predetermined growth direction method developed in [41, 42, 46].
[17] Werner P, Zakharov N D, Gerth G, Schubert L, Sokolov L and G¨osele U 2006 J. Cryst. Growth 290 6 [18] Ohlsson B J, Bj¨ork M T, Magnusson M H, Deppert K, Samuelson L and Wallenberg L R 2001 Appl. Phys. Lett. 79 3335 [19] Hiruma K, Yazawa M, Haraguchi K, Ogawa K, Katsuyama T, Koguchi M and Kakibayashi H 1993 J. Appl. Phys. 74 3162 [20] Dubrovskii V G, Soshnikov I P, Sibirev N V, Cirlin G E and Ustinov V M 2006 J. Cryst. Growth 289 31 [21] Duan X, Yu H, Yl C, Jianfang W and Lieber C M 2001 Nature 409 66 [22] Kar S, Biswas S and Chaudhuri S 2005 Nanotechnology 16 737 [23] Zhang X T, Liu Z, Leung Y P, Li Q and Hark S K 2003 Appl. Phys. Lett. 83 5533 [24] Colli A, Hofmann S, Ferrari A C, Martelli F, Rubini S, Ducati C, Franciosi A and Robertson J 2005 Nanotechnology 16 S139 [25] Sochinskii N V, Silveira J P, Briones F, Saucedo E, Herrero C M, Fornaro L, Bermudez V and Dieguez E 2005 J. Cryst. Growth 275 e1131 [26] Krahne R, Chilla G, Schuller C, Carbone L, Kudera S, Mannarini G, Manna L, Heitmann D and Cingolani R 2006 Nano Lett. 6 478 [27] Li L, Yang Y, Huang X, Li G and Zhang L 2005 J. Phys. Chem. B 109 12394 [28] Huo H B, Dai L, Xia D Y, Ran G Z, You L P, Zhang B R and Qin G G 2006 J. Nanosci. Nanotechnol. 6 1182 [29] Huo H B, Dai L, Liu C, You L P, Yang W Q, Ma R M, Ran G Z and Qin G G 2006 Nanotechnology 17 5912 [30] Wagner R S and Ellis W C 1964 Appl. Phys. Lett. 4 89 [31] Dubrovskii V G, Cirlin G E, Soshnikov I P, Tonkikh A A, Sibirev N V, Samsonenko Y and Ustinov V M 2005 Phys. Rev. B 71 205325 [32] Dubrovskii V G, Sibirev N V, Cirlin G E, Harmand J C and Ustinov V M 2006 Phys. Rev. E 73 021603 [33] Johansson J, Svensson C P T, Martensson T, Samuelson L and Seifert W 2005 J. Phys. Chem. B 109 13567 [34] Furdyna J K 1988 J. Appl. Phys. 64 R29 [35] Kulkarni J S, Kazakova O and Holmes J D 2006 Appl. Phys. A 85 277 [36] Wolf S A, Awschalom D D, Buhrman R A, Daughton J M, von Molnar S, Roukes M L, Chtchelkanova A Y and Treger D M 2001 Science 294 1488 [37] Janik E et al 2006 Appl. Phys. Lett. 89 133114 [38] Duan X, Jianfang W and Lieber C M 2000 Appl. Phys. Lett. 76 1116 [39] Hiruma K, Yazawa M, Katsuyama T, Ogawa K, Haraguchi K, Koguchi M and Kakibayashi H 1995 J. Appl. Phys. 77 447 [40] Wu Z H, Mei X, Kim D, Blumin M, Ruda H E, Liu J Q and Kavanagh K L 2003 Appl. Phys. Lett. 83 3368 [41] Chan S K, Cai Y, Wang N and Sou I K 2006 Appl. Phys. Lett. 88 013108 [42] Cai Y, Chan S K, Sou I K, Chan Y F, Su D S and Wang N 2005 Adv. Mater. 18 109 [43] Krishnamachari U, Borgstrom M, Ohlsson B J, Panev N, Samuelson L, Seifert W, Larsson M W and Wallenberg L R 2004 Appl. Phys. Lett. 85 2077 [44] Jensen L E, Bj¨ork M T, Jeppesen S, Persson A I, Ohlsson B J and Samuelson L 2004 Nano Lett. 4 1961 [45] Wang H and Fischman G S 1994 J. Appl. Phys. 76 1557 [46] Chan S K, Cai Y, Wang N and Sou I K 2007 J. Cryst. Growth 301/302 866
Acknowledgments The authors would like to thank Dr S Ma´ckowski and Professor H P Strunk for fruitful discussions and careful reading of the manuscript. This research was partially supported by the Ministry of Science and Higher Education (Poland) through Grants N507 030 31/0735 and N515 015 32/0997, and by the Network ‘New materials and sensors for optoelectronics, information technology, energetic applications and medicine’.
References [1] Samuelson L et al 2004 Physica E 25 313 [2] Gudiksen M S, Lauhon L J, Wang J, Smith D C and Lieber C M 2002 Nature 415 617 [3] Seifert W et al 2004 J. Cryst. Growth 272 211 [4] Dubrovskii V G and Sibirev N V 2004 Phys. Rev. E 70 031604 [5] Cui Y and Lieber C M 2001 Science 291 851 [6] Huang Y, Duan X, Cui Y, Lauhon L J, Kim K H and Lieber C M 2001 Science 294 1313 [7] Samuelson L 2003 Mater. Today 6 22 [8] Cui Y, Wei Q, Park H and Lieber C M 2001 Science 293 1289 [9] Qian F, Li Y, Gradecak S, Wang D, Barrelet C J and Lieber C M 2004 Nano Lett. 4 1975 [10] Duan X, Yu H, Agarwal R and Lieber C M 2003 Nature 421 241 [11] Bj¨ork M T, Ohlsson B J, Thelander C, Persson A I, Deppert K, Wallenberg L R and Samuelson L 2002 Appl. Phys. Lett. 81 4458 [12] Thelander C, M˚artensson T, Bj¨ork M T, Ohlsson B J, Larsson M W, Wallenberg L R and Samuelson L 2003 Appl. Phys. Lett. 83 2052 [13] Wang J, Gudiksen M S, Duan X, Cui Y and Lieber C M 2001 Science 293 1455 [14] Schubert L, Werner P, Zakharov N D, Gerth G, Kolb F M, Long L, G¨osele U and Tan T Y 2004 Appl. Phys. Lett. 84 4968 [15] Westwater J, Gosain D P, Tomiya S, Usui S and Ruda H 1997 J. Vac. Sci. Technol. B 15 554 [16] Kodambaka S, Tersoff J, Reuter M C and Ross F M 2006 Phys. Rev. Lett. 96 096105
8
