Defect-Induced Nucleation and Epitaxy: A New Strategy toward the

Oct 30, 2015 - In this work, we demonstrate a new strategy to create WZ-GaN/3C-SiC heterostructure nanowires, which feature controllable morphologies...
2 downloads 0 Views 2MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Defect-Induced Nucleation and Epitaxy: A New Strategy toward the Rational Synthesis of WZ-GaN/3C-SiC Core−Shell Heterostructures Baodan Liu,*,† Bing Yang,† Fang Yuan,† Qingyun Liu,† Dan Shi,† Chunhai Jiang,‡ Jinsong Zhang,‡ Thorsten Staedler,§ and Xin Jiang*,†,§ †

Shenyang National Laboratory for Materials Science (SYNL) and ‡Materials Fabrication and Processing Division, Institute of Metal Research (IMR), Chinese Academy of Sciences (CAS), No. 72 Wenhua Road, Shenyang 110016 China § Institute of Materials Engineering, University of Siegen, Germany, Paul-Bonatz-Strasse 9−11, 57076 Siegen, Germany S Supporting Information *

ABSTRACT: In this work, we demonstrate a new strategy to create WZ-GaN/3CSiC heterostructure nanowires, which feature controllable morphologies. The latter is realized by exploiting the stacking faults in 3C-SiC as preferential nucleation sites for the growth of WZ-GaN. Initially, cubic SiC nanowires with an average diameter of ∼100 nm, which display periodic stacking fault sections, are synthesized in a chemical vapor deposition (CVD) process to serve as the core of the heterostructure. Subsequently, hexagonal wurtzite-type GaN shells with different shapes are grown on the surface of 3C-SiC wire core. In this context, it is possible to obtain two types of WZ-GaN/3C-SiC heterostructure nanowires by means of carefully controlling the corresponding CVD reactions. Here, the stacking faults, initially formed in 3C-SiC nanowires, play a key role in guiding the epitaxial growth of WZ-GaN as they represent surface areas of the 3C-SiC nanowires that feature a higher surface energy. A dedicated structural analysis of the interfacial region by means of high-resolution transmission electron microscopy (HRTEM) revealed that the disordering of the atom arrangements in the SiC defect area promotes a lattice-matching with respect to the WZ-GaN phase, which results in a preferential nucleation. All WZ-GaN crystal domains exhibit an epitaxial growth on 3C-SiC featuring a crystallographic relationship of [12̅10]WZ‑GaN //[011̅]3C‑SiC, (0001)WZ‑GaN//(111)3C‑SiC, and dWZ‑GaN(0001) ≈ 2d3C‑SiC(111). The approach to utilize structural defects of a nanowire core to induce a preferential nucleation of foreign shells generally opens up a number of opportunities for the epitaxial growth of a wide range of semiconductor nanostructures which are otherwise impossible to acquire. Consequently, this concept possesses tremendous potential for the applications of semiconductor heterostructures in various fields such as optics, electrics, electronics, and photocatalysis for energy harvesting and environment processing. KEYWORDS: Wurtzite GaN, 3C-SiC, core−shell heterostructure, stacking faults, confined epitaxial growth

B

constants and the same crystallographic geometry. In such a case, the components usually display the same crystalline structure and either share the same constituent element or feature constituent elements from the same groups in order to show related physical and chemical properties. Representative examples of such heterostructures can be found in numerous binary and ternary systems such as Si/Ge,2,5,9 Si/GaP,10 ZnO/ ZnS,11−14 GaP/InP,15 GaAs/InAs,15,16 GaP/GaAs,4 InP/ InAs,6,17,18 GaN/InGaN,3,19,20 AlN/GaN,21 GaN/AlGaN,22 GaAsP/GaP, GaN/AlN/AlGaN,23 and GaN/ZnO.7 On the other hand, a heterostructure of two materials, which possess different crystallographic structures, is also possible but only if they feature similar crystal lattice parameters in the vicinity of their interface in order to achieve a good lattice-matching. A

ecause of their intriguing properties and promising applications in electronics, optoelectronics, and photocatalysis, low-dimensional heterostructure nanowires comprised of different semiconductor components have been the subject of a remarkable fundamental as well as technological research interests in past years.1−7 The actual combination of different semiconductors at their phase boundary, which allows for the realization of unique functions, requires a careful growth design. A number of representative examples of semiconductor heterostructures are reported in the literature. Among those, Si/Ge, ZnO/GaN, and ZnO/Si, which are employed in the context of specific applications such as field-emission transistor (FET),5 light-emitting diode (LED),7 and photovoltaic solarcell and H2-generation photocatalyst, respectively, are typical candidates.8 Furthermore, low-dimensional heterostructures are generally classified into three categories, which are referred to as axial, radial, and a core−shell one. With respect to a minimized interfacial energy, a heterojunction is normally composed of two materials, which feature similar lattice © XXXX American Chemical Society

Received: June 20, 2015 Revised: October 20, 2015

A

DOI: 10.1021/acs.nanolett.5b02454 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

core−shell ones. The stacking fault section initially formed in a 3C-SiC nanowire can precisely confine the epitaxial nucleation and first growth of the separated GaN shells. We also demonstrate that the shell morphology of WZ-GaN/3C-SiC heterostructures can be selectively modified from periodic GaN hexagonal truncated pyramids to continuous GaN shells through a control of the reaction time and two types of WZGaN/3C-SiC heterostructures have been prepared. The morphology, microstructure, crystallography and local composition of the samples produced within the framework of this study are rigorously analyzed. The formation mechanism and growth process of WZ-GaN/3C-SiC heterostructures are discussed based on detailed structural and compositional results. To synthesize WZ-GaN/3C-SiC heterostructures a two-step CVD process was designed to first produce the 3C-SiC nanowire cores and, subsequently, the WZ-GaN shells. It is schematically described in Supporting Information Figure S1. Initially, the 3C-SiC nanowires are synthesized by the carbonization of Si powders at a reaction temperature of 1400 °C in a horizontal resistance furnace.37 The carbonization of Si vapor at high temperature produces a high-yield of SiC nanowires with an average diameter of ∼100 nm. Scanning electron microscopy (SEM) reveals that the as-synthesized SiC nanowires are straight and free of metal catalyst particles at the tip-ends (Supporting Information Figure S2), suggesting that a direct vapor−solid process is involved in their nucleation. The as-synthesized SiC nanowires were afterward transferred to a second CVD reactor for the deposition of the corresponding GaN shells.38−40 In order to promote a uniform deposition of GaN on the SiC nanowires, the loose cotton-like SiC nanowires were directly placed on the surface of a Ga2O3 powder source. Figure 1 schematically illustrates the formation process of the

typical example of such heterostructure can be found in Si/ ZnSe axial heterostructure nanowires. Here the (001) plane of the cubic Si matches well with the (011̅0) plane of the hexagonal ZnSe at the domain boundary.24 For low-dimensional heterostructure nanowires composed of components with different geometrical symmetry and chemical composition, the controlled growth of a core−shell morphology is extremely challenging in comparison to its axial and radial counterparts and has rarely been reported. To obtain heterostructure nanowires, several relative simple easy to control strategies have been developed in the past years. For instance, heterostructure nanowires with an alternating composition along the growth direction can be fabricated by periodically changing the precursor transportation in a chemical vapor deposition (CVD) or plasma-assisted molecular-beam epitaxy (PAMBE) process.1,21 Similar heterojunctions, which feature a core−shell configuration, can be obtained exploiting an analogous lateral epitaxy strategy.7,25 Recently, branched heterostructure nanoarchitectures comprising of different materials have attracted intensive research interest. These structures typically are fabricated by multistep CVD, MBE, or solution processes.15,26,27 Considering the significance of each segment in influencing the heterostructure properties, accurate control of size, morphology, and fraction of each component as well as the corresponding interface, is of extreme importance. So far, quantum dot heterostructures with precise section length have been achieved in some III−V heterostructure nanowires by modulating the precursor in a MOCVD or MBE process.20 However, the degree of control required for such a process still renders the analogous epitaxial growth of core− shell heterostructure nanowires challenging. A good example in this context is the option of forming periodic separated shells along a nanowire core. The latter still poses a significant challenge despite substantial efforts that have been undertaken.28 Inspired by the defect-driven preferential nucleation of numerous nanostructures29−32 in which the nucleation sites can be precisely confined by corresponding defect positions, we presumed that a nanowire containing structural defect such as stacking faults will offer opportunities for a selective nucleation of outer shells. Such consideration implies that shell size and position can be determined by the defect sections in the nanowire cores. As a result, this would allow for the controlled synthesis of core−shell heterostructure nanowires with separated shells. Herein, we report the rational fabrication of WZ-GaN/3CSiC heterostructures by means of a well-designed two-step CVD process. The cubic 3C-SiC nanowires with periodic stacking fault sections are employed as cores, while the WZGaN shells with a controllable size are subsequently grown by epitaxy. Generally, the [0001]-oriented 6H-SiC crystalline substrate with wurtzite-type structure was the first choice for an epitaxial growth of WZ-GaN crystals due to their resemblance in lattice constants.33,34 The smaller difference of lattice constants between 6H-SiC and WZ-GaN allows for a decent lattice-matching at the SiC/GaN interface and thus for obtaining highly crystallized GaN epitaxial layer with less structural strain. In comparison with 6H-SiC, cubic 3C-SiC is rarely used for the epitaxial growth of WZ-GaN layer due to its relative lattice mismatching with WZ-GaN.35,36 Consequently, up to now the path via 6H-SiC is generally favored in order to achieve an epitaxial growth of a GaN layer.35,36 However, we demonstrate that the formation mechanism of WZ-GaN/3CSiC heterostructures is totally different from other axial or

Figure 1. Schematic diagram describing the formation process of the WZ-GaN/3C-SiC core−shell heterostructure nanowires.

WZ-GaN/3C-SiC heterostructures starting from the initial 3CSiC nanowires and ending with continuous core/shell structures. In a typical growth process, the nucleation of WZGaN shells on the 3C-SiC nanowires is initially confined by the stacking fault sections of the cores. These sections feature a higher surface energy therefore enabling a preferential absorption of GaN nanocrystals (step I in Figure 1). For a prolonged reaction time, these GaN nanoislands further evolve into hexagonal truncated pyramid-like shells with crystalline facets by lateral epitaxial growth (step II in Figure 1). Any further extension of the reaction time will result in the consecutive crystallization of GaN shells at their bottoms leading to the formation of continuous GaN shells (step III in Figure 1). As the core of the WZ-GaN/3C-SiC heterostructure, the 3CSiC nanowire plays a key role in the initial nucleation and B

DOI: 10.1021/acs.nanolett.5b02454 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 2. (a) STEM image of 3C-SiC nanowire with periodic stacking fault sections and (b,c) magnified STEM images of stacking fault areas; (d) atomically resolved lattice image of 3C-SiC nanowire free of stacking fault and (e) its corresponding SAED pattern; (f) interface HRTEM image of well-crystallized 3C-SiC crystal domain and stacking fault area.

Table 1. Structure Parameters of 3C-SiC, Hexagonal SiC, ZB-GaN, and WZ-GaN SiC polymorphs

Pdf No

3C-SiC 2H-SiC 4H-SiC 6H-SiC 8H-SiC

29-1129 29-1130 29-1127 29-1131 48-0708

GaN lattice constants (nm) a a a a a

= = = = =

0.436 0.308, 0.308, 0.307, 0.308,

c c c c

= = = =

0.505 1.006 1.508 2.014

epitaxial growth of the WZ-GaN shells. The details of a 3C-SiC nanowire are presented by a representative bright-field TEM image (Figure 2a). Here periodic dark sections, alternately separated by bright areas, are observed within the wire. These dark sections with a length of 100−300 nm correspond to stacking fault areas induced by a disordering stacking sequence of atoms (Figure 2b), which is characteristic for 3C-SiC nanowires.41 The bright areas with comparable size as the dark ones are considered as well-crystallized crystal domains of 3CSiC even though occasionally some stacking faults along the (111) plane can be observed (Figure 2c).42 From Figure 2a, it can be seen that the length of defect areas is extremely difficult

polymorphs

Pdf No

lattice constants (nm)

ZB-GaN

52-0871

a = 0.45

WZ-GaN

74-0243

a = 0.318 c = 0.517

to control due to the self-organized formation process. Highresolution TEM (HRTEM) analysis at the edge of such bright sections reveals a regular arrangement of atoms in a facecentered cubic (fcc) structure, indicating that the as-synthesized SiC nanowires are indeed cubic 3C-SiC wires.43,44 Selected area electron diffraction (SAED) pattern taken along the [110] zone axis from these well-crystallized section also verifies the good crystallinity and crystallographic structure of the SiC nanowires (Figure 2d,e). The structural data of the 3C-SiC wires are further supported by the X-ray diffraction (XRD) results, which show a typical pattern of a zinc-blende-type fcc structure (Supporting Information Figure S3). However, it is noted that C

DOI: 10.1021/acs.nanolett.5b02454 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

the generation of numerous atomic steps and increasing the surface areas of the corresponding length, which has relatively higher total surface energy for a preferential nucleation and formation of GaN nanocrystals on the surface. Composition analysis utilizing an X-ray energy dispersive spectrometer (EDS) technique confirms that these nanoislands are made of Ga and N elements with a stoichiometric ratio approaching to Ga/N = 1:1, suggesting the formation of GaN on 3C-SiC surface. From the magnified TEM image shown in Figure 3b, one notices that all the GaN nanoislands are regularly nucleated on the protruding surface of the 3C-SiC nanowires in a discontinuous circular fashion, which implies that the GaN nanocrystals exhibit a preferential nucleation even for the stacking fault. HRTEM analysis performed on 3C-SiC nanowires demonstrates that the wires are oriented along the [111] direction (Figure 2d,f). Additionally, an expansion in diameter of the wire correlated to the stacking fault sections is observed, which is readily explained by the change in stacking sequence. As a result, the defect area surface will become more energy-favorable for the initial nucleation of GaN nanoislands, as schematically illustrated in Figure 3c. The TEM result shown in Figure 3a demonstrated the possibility of GaN crystal growth on a 3C-SiC surface, which features periodic stacking fault areas. These GaN nanoislands initially formed on the surface of the SiC nanowires can serve as seeds, which guide a subsequent growth of GaN shells. Therefore, it seems reasonable that a prolonged reaction time will result in a continuous GaN shells with large grain size. To verify this assumption, the depositions of GaN on SiC nanowires was carried out for 60 and 120 min, respectively, while all other growth parameters were kept constant. Figure 4

the (220) peak in the XRD pattern shows stronger intensity in comparison to the (111) peak. This contradicts the statement for a standard XRD pattern of 3C-SiC reported in pdf card (No. 29-1129), which implies that the (111) crystalline plane should show the strongest intensity. The changed experimental finding in this study arises from the fact that the SiC nanowires contain a large amount of stacking faults, which derive from the variation of atom stacking sequence. These stacking faults may induce the appearance of a hexagonal SiC phase with different polymorphs, while at the same time all these hexagonal SiC phases share the same interplanar distance with the (220) plane of the 3C-SiC (Supporting Information Figure S3). In this context, more than 150 polymorphs of hexagonal SiC are reported with respect to the cubic 3C-SiC. All of these SiC polymorphs featuring different crystallographic symmetries can be reversely transformed through a variation of the stacking sequence.45 Table 1 lists the main lattice parameters of hexagonal SiC and 3C-SiC. Remarkably, all hexagonal SiC only differ slightly in their c-plane distance. Consequently, the observed difference in diffraction intensity between the (111) and the (220) peaks of the XRD pattern appears plausible. Figure 2f shows a HRTEM image of the interface region between a stacking fault section and its adjacent crystalline 3CSiC section of the nanowires. It is found that the crystalline boundary between the two neighboring crystal domains is atomically flat. The bright area (left part in Figure 2f) exhibits a superior crystallinity with all atoms well organized, while the dark section is full of high-density structural defects comprised of different polymorphs of hexagonal SiC. Figure 3 shows a typical TEM image of a 3C-SiC nanowires decorated with some small nanoislands. These islands, which

Figure 4. (a,c) Low-magnification and (b,d) high-resolution SEM images of WZ-GaN/3C-SiC heterostructure nanowires grown at 1100 °C for 60 and 120 min, respectively, revealing the morphology evolution of GaN hexagonal truncated pyramids to continuous shells with the prolongation of reaction time.

Figure 3. (a) Representative TEM image of WZ-GaN/3C-SiC heterostructure nanowire grown at 1100 °C for 5 min; (b) magnified TEM image of the stacking fault area, revealing the aggregation of GaN nanoislands preferentially nucleated on the 3C-SiC nanowire surface; (c) crystallographic model for illustrating the favored nucleation of GaN nanoislands on the protruding (111) planes induced by the stacking faults.

shows typical morphologies of GaN/3C-SiC heterostructures grown for different reaction times. Notably, an extension of the growth time to 60 min leads to a change in morphology. The small nanoislands evolve into hexagonal truncated pyramids with a maximum bottom size of 200−300 nm (Figure 4a,b). The symmetrical 6-fold geometry of GaN crystals with smooth and apparent crystal facets suggests that crystalline WZ-GaN domains are developed from the tiny GaN seeds. This process is schematically described in Figure 1. At this stage, the alternating stacking fault sections of the 3C-SiC nanowires are

show a size of several nanometers and an irregular shape, are all situated on corresponding stacking fault sections after a reaction at 1100 °C for 5 min. Interestingly, no nanoislands are found on the well-crystallized regions of the 3C-SiC nanowires. The disordered stacking of atoms in a 3C-SiC nanowires will locally alter their stacking sequences, leading to D

DOI: 10.1021/acs.nanolett.5b02454 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 5. (a,b) Typical STEM images of WZ-GaN3C-/SiC heterostructure nanowire grown for 60 and 120 min, respectively; (c−e) atomically resolved HRTEM images of GaN shells and WZ-GaN/3C-SiC overlapping area; (f−h) SAED patterns recorded from GaN shell and WZ-GaN/3CSiC overlapping areas. The diffraction lines in (f) verifies the high density of stacking faults in GaN shells; whereas the diffraction spots from different SAED patterns justify the existence of WZ-GaN and 3C-SiC phases.

ture nanowires with separated and continuous GaN shells, respectively. The evolution of the GaN shell morphology from a hexagonal truncated pyramid one to linked layers with gradient diameters corresponding to the growth time is identical with the one shown in Figure 4. HRTEM analysis performed on the side wall of a hexagonal truncated GaN pyramid provides images showing clear atomic fringes that indicate good ordering (Figure 5c). On the surface edge, the GaN crystal shows a superior crystallinity while the inner part contains a high density of structural defects (stacking faults, Figure 5c). The d-spacing of 0.51 and 0.27 nm measured from the neighboring lattice planes along the longitudinal and lateral direction of the heterostructure nanowire, respectively, match well with the interplanar distances of the (0001) and (101̅0) planes of bulk WZ-GaN crystal. This finding strongly indicates that the outer shell features the WZ-GaN phase. A corresponding SAED pattern taken from the inner part of the hexagonal truncated pyramids, shown in Figure 5f, displays well-arranged spots linked with thin lines (marked by arrows). All spots that show strong diffraction intensity can readily be indexed to the corresponding crystalline planes of WZ-GaN. The thin lines marked with arrows, however, reflect the formation of a high density of stacking faults in GaN crystals, which are induced by the initial stacking faults of the 3C-SiC nanowire cores during the growth process. To shed some light on the confined epitaxial growth of hexagonal truncated GaN pyramids (Figure 5a), the GaN deposition on SiC nanowire cores is studied. Figure 5d shows a typical HRTEM image of an overlapping region of the WZ-GaN/3C-SiC structure. Three characteristic domains can be identified: area I for a pure WZGaN shell; area II for a WZ-GaN/3C-SiC core−shell structure; and area III for the original well-crystallized 3C-SiC nanowire. The stacking faults can not only be observed in area II but also

completely covered by the faceted WZ-GaN crystals, while the well-crystallized regions of the wires still remain uncovered. Interestingly, the height of all the WZ-GaN hexagonal truncated pyramids along the 3C-SiC nanowire core is precisely confined by the length of the stacking fault areas. Spatially resolved elemental mapping performed on the WZ-GaN/3CSiC heterostructure nanowires further verifies the defectinduced epitaxial growth of WZ-GaN on the 3C-SiC surface along with its precise size confinement (Supporting Information Figure S4). In the case of a reaction time of 120 min, consecutive GaN shells with a gradient thickness are deposited on the 3C-SiC nanowire surface to form coaxial WZ-GaN/3CSiC heterostructure nanowires (Figure 4c,d). EDS line-scan analysis along the lateral direction clearly confirms the core− shell feature (Supporting Information Figure S5). Notably, the later grown GaN shells exhibit a totally different configuration and crystallization process with respect to the GaN hexagonal truncated pyramids shown in Figure 4a,b. These GaN crystals start to crystallize from the bottom of the hexagonal truncated pyramid, which could be directly observed in the TEM image shown in Supporting Information Figure S6 (and is in Figure S6b, in which the yellow dash line areas represent the preformed hexagonal truncated pyramid, whereas the white dash line areas correspond to the newly grown GaN crystal from the bottom). As a key semiconductor material, GaN generally exists in its energetically stable wurtzite-type hexagonal or metastable zinc-blende-type cubic configurations, respectively (Table 1).46,47 To get detailed information about the crystallographic structure of the WZ-GaN shells and their crystallinity, the WZ-GaN/3C-SiC heterostructure nanowires grown for 60 and 120 min were investigated by HRTEM. Figure 5a,b shows representative TEM images of WZ-GaN/3C-SiC heterostrucE

DOI: 10.1021/acs.nanolett.5b02454 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 6. (a) Interface HRTEM image of GaN/SiC heterostructure nanowire revealing the crystallographic relationship between GaN and SiC; (b) crystal model describing the bonding of GaN with SiC at different crystal domain boundaries.

in area I, the pure GaN shell. This finding suggests that the stacking faults of the 3C-SiC nanowire core propagate into the WZ-GaN shell, while the sharp boundary between the wellcrystallized 3C-SiC and WZ-GaN/3C-SiC heterostructure again verifies the precisely confined nucleation/crystallization of WZ-GaN shell by means of the stacking fault defect of the 3C-SiC nanowires. It should be pointed out that the stacking faults of the 3C-SiC nanowire would only penetrate into the very first GaN layers in case of a preferential nucleation on the 3C-SiC defect areas. A growing GaN shell, however, should result in an improvement of crystallinity based on a structural strain release. As a result, crystalline domains free of structural defect would be obtained. In the 120 min deposition case, WZGaN shells with faceted crystal growth fronts corresponding to the (0001) and (101̅1) planes of WZ-GaN can be observed (Figure 5e). The angle of 118° between the two neighboring facets agrees reasonably well with the theoretical angle included between the two planes, while the succinct lattice fringe image confirms the superior crystallinity at the edge (Figure 5e). The two representative SAED patterns shown in Figure 5g,h reveal that the overlapping WZ-GaN/3C-SiC volume yields at least three distinctive patterns. Additionally, thin diffraction lines corresponding to stacking faults are observed. For example, the SAED patterns in Figure 5g can be divided into three separated diffraction patterns with respective crystalline units. The diffraction spots connected by the red lines can be indexed to the (0001) and (101̅0) planes of WZ-GaN along a zone axis of [121̅ 0]; while the other spots linked with yellow and cyan lines correspond to the bicrystalline twins of the 3C-SiC nanowires in which the two group planes of (200)/(200)T and (022)/ (022)T are symmetrically mirrored by the crystalline twin boundary of the (222) plane. From such a SAED pattern it is possible to deduce the crystallographic relationship between WZ-GaN and 3C-SiC. The (0002) plane of WZ-GaN matches well with the (111) plane of 3C-SiC. Unfortunately, the atomically resolved lattice fringe of this area can not be obtained due to the huge thickness of the overlapped crystals. Figure 5h represents yet another SAED pattern taken from the heterostructure nanowires corresponding to 120 min growth. Similar to the pattern shown in Figure 5g, at least three individual SAED patterns corresponding to WZ-GaN and 3CSiC can be distinguished. The appearance of multiset diffraction spots, however, implies that extra GaN grains are formed during the morphological evolution of the GaN shells, similar to the case of In2O3/SnO2 core−shell heterostructure nanowires.48

Further dedicated TEM investigations on the interface between the GaN shell and the SiC nanowire core are pursued in order to explore the initial preferential nucleation of WZGaN domains on the stacking fault surface area of the 3C-SiC nanowire as well as the evolution of the GaN crystallization itself. Figure 6a presents an atomic resolution TEM image of a WZ-GaN shell and the corresponding 3C-SiC core. Because of the large thickness and high density of structural defects, the phase (or crystal domain) boundaries among the four feature areas can be roughly identified, as separated by the green (phase boundary) and yellow (crystallographic symmetry boundary) dashed lines. Area (1) represents the well-crystallized 3C-SiC nanowire core. The atom stacking order observed here is consistent with a sequence of ABCABC (yellow circles). The area defined as (2) protruding from the 3C-SiC core surface corresponds to the epitaxial GaN nucleation, which still shows the same atom stacking order as the corresponding 3CSiC core, a sequence of ABCABC (green circles). The same lattice configuration found for area (2) and area (1) verifies the ZB-GaN phase in the GaN shell. The excellent lattice-matching between ZB-GaN and 3C-SiC gives proof for the confined local epitaxial growth of ZB-GaN on the 3C-SiC surface with a crystalline orientation relationship of [11̅0]ZB‑GaN//[11̅0]3C‑SiC, (111)ZB‑GaN//(111)3C‑SiC. The thickness of the ZB-GaN phase is roughly estimated to be only several atomic layers. Area (3) corresponds to the central part of WZ-GaN shell. As it approaches the center of the stacking fault area, the atom stacking sequence in GaN shell exhibits a tendency of disordering and obvious stacking faults can be observed. Correspondingly, the GaN shell is mainly composed of WZGaN featuring a relative poor crystallinity (area (3)). Additionally, this result is confirmed in Figure 5d. Area (4) reveals the overlapping region of the WZ-GaN/3C-SiC core− shell heterostructure. Here, a high density of structural defects can be found. It is commonly known that SiC features numerous polymorphs including cubic 3C-SiC as well as hexagonal nH-SiC with n = 2, 4, 6, and so forth (Supporting Information Figure S3 and Table 1).49−51 Although the transition of 3C-SiC to hexagonal SiC holds many possibilities for the atom stacking sequence along the c-axis, which allows for a formation of SiC polymorphs with variable lattice distances along the [0001] direction, the distance between the (0001) planes of WZ-GaN is fixed. As a consequence, the (0001) plane of WZ-GaN presumably combines with the (0001) plane of various hexagonal SiC polymorphs that in turn leads to the complicated lattice-matching within the stacking F

DOI: 10.1021/acs.nanolett.5b02454 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 7. (a) Schematic model of WZ-GaN/3C-SiC heterostructure nanowire with continuous GaN shells; (b) a typical low-magnification TEM image of WZ-GaN/3C-SiC heterostructure nanowire grown for 120 min; (c,d) high-magnification TEM images of the selected areas (“a” and “b”) in a; (e−g) atomically resolved HRTEM images of the selected areas in the interface of WZ-GaN/3C-SiC heterostructures.

of WZ-GaN on a 3C-SiC nanowire surface implies that the hexagonal WZ-GaN may have some crystallographic orientation relationship with the cubic 3C-SiC induced by its initial nucleation on a stacking fault area of the 3C-SiC and subsequent crystallization. To explore the possibility of an epitaxial growth of WZ-GaN crystal domains and its lattice matching with respect to the 3C-SiC nanowires the WZ-GaN/ 3C-SiC heterostructure nanowires grown for 120 min were analyzed by HRTEM. Figure 7a,b shows a schematic diagram and a TEM image of a WZ-GaN/3C-SiC nanowire with a WZGaN shell that covers the majority of the 3C-SiC nanowire. Typically, outer WZ-GaN shells with irregular shapes still possess some facets obviously featuring a wurtzite-type of structure. Additionally, bare SiC nanowire regions are observed occasionally. Figure 7b shows two GaN crystal domains deposited on a 3C-SiC surface, which almost reached coalescence (marked by the arrows). On the basis of the SEM and TEM analyses shown above (Figures 4 and 5), the bare 3C-SiC nanowires should display a fcc-type cubic structure free of structural defects while the outer GaN shell favors a wurtzite-type hexagonal structure. The areas for which the SiC nanowire is covered by a thin GaN shell represents the ideal locations for a systemic investigation of the structural relationship between WZ-GaN and 3C-SiC and deduces how WZ-GaN grows on a 3C-SiC nanowire surface. Figure 7c,d presents HRTEM images of the two representative areas marked in Figure 7b (area “a” shows a typical region for the coalescence of two individual GaN crystal domains; whereas area “b” represents a defect-induced epitaxial growth of GaN shell on SiC nanowire). The crystal domain boundaries of the different grains can easily be identified and are marked by dashed lines. In Figure 7c, two WZ-GaN crystal grains with different orientations are observed, which are tightly attached to the surface of the SiC nanowire core. The white dashed line represents the boundary of two WZ-GaN crystal domains. Interestingly, the two WZ-GaN grains (shells) exhibit an

fault area. From the HRTEM image shown in Figure 6a, a lattice distortion occurring at the WZ-GaN/3C-SiC interface due to the lattice difference of GaN and SiC as well as the evolution of geometrical structures is observed and indicated by red arrows. For perfect ZB-GaN and 3C-SiC crystals, the lattice constants match so well (0.446 nm for ZB-GaN (JPCDS No. 88-2364) and 0.436 nm for 3C-SiC (JCPDS No. 29-1129), Table 1) that their difference is insufficient to induce any significant lattice distortion. However, the disorder of the atom stacking sequence induced by the stacking fault in the 3C-SiC core and the WZ-GaN shell will produce large structural strain and lead to a variation of the GaN fringe. As a result, an obvious lattice mismatch between SiC and GaN is formed to ensure structural stability. To further elucidate the crystallographic relationship between SiC and GaN with respect to their different structures and symmetries, a crystal model comprised of perfect cubic 3C-SiC and ZB-GaN, hexagonal SiC, and WZGaN crystal domains is set up and illustrated in Figure 6b. This representation significantly facilitates the understanding of the atomic stacking sequence in cubic SiC (or GaN) and hexagonal SiC (or GaN) as well as the lattice matching along different crystalline boundaries. Clearly, the chemical bonding and lattice matching along the crystal domain boundaries exhibit distinct differences with respect to the individual crystal domains resulting from the structure and phase transition. In order to maintain the peculiar heterostructure crystallographically stable, all atoms at the interface will deviate from their original position, which leads to a structural distortion that can be observed by HRTEM imaging. The results presented above clearly demonstrated the preferential nucleation and epitaxial growth of WZ-GaN crystal domains on the surface of 3C-SiC nanowires based on stacking faults. However, it is found that a WZ-GaN deposition is also possible in the case of a defect free 3C-SiC surface after the preferential nucleation and crystallization on the stacking fault sections took place (Figures 4c,d and 5b). A complete wrapping G

DOI: 10.1021/acs.nanolett.5b02454 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Conclusion. In summary, we reported a new strategy for the selective nucleation of WZ-GaN on cubic 3C-SiC nanowires to form WZ-GaN/3C-SiC heterostructures. The latter are promising candidates for numerous optoelectronic applications. The stacking faults initially formed in the 3C-SiC nanowires promote the preferential nucleation of GaN nanoislands on the SiC surfaces as a result of their higher surface energy. The morphology and interface of the WZ-GaN/ 3C-SiC heterostructures can easily be tailored by controlling the reaction duration. A morphological evolution from a periodic hexagonal truncated GaN pyramids case to a continuous GaN shell with gradient thickness has been realized. The utilization of structural defects as the preferential nucleation sites for the heteroepitaxial growth of WZ-GaN on 3C-SiC nanowires overcomes the hurdles of conventional methods in achieving a direct epitaxy of WZ-GaN on a 3C-SiC substrate. Such a WZ-GaN/3C-SiC heterostructure system with controllable morphology and a designable interface allows for a selective band gap engineering and efficient electron−hole separation, which, in turn leads to promising optoelectronic applications. Exactly such type of investigation, namely the evaluation of the WZ-GaN/3C-SiC heterostructure nanowires for use as high-speed field-effect transistor (FET) or efficient photocatalysts for water-splitting (production of hydrogen), is currently a work in progress. Methods. Nanowire Synthesis. SiC nanowires with periodic size variation and bamboo-like morphology were synthesized in a horizontal resistance furnace by means of a facile CVD process. High-purity carbon and SiO2 powders mechanically mixed and loaded on an Al2O3 crucible were used as the precursors of C and Si, respectively. The crucible was put into the center of a Al2O3 tube inside the furnace and then heated to 1400 °C under a stable Ar flowing rate of 200 mL/ min for promoting the chemical reaction of the SiC nanowire formation. After growing at this temperature for 30 min, a large yield of cotton-like SiC nanowires with transparent green color densely cover the whole surface of the Al2O3 crucible. To achieve WZ-GaN/3C-SiC heterostructures, the as-grown 3CSiC nanowires removed from the Al2O3 crucible surface were transferred to the furnace for GaN synthesis. The 3C-SiC nanowires were directly put on the top of Ga2O3 powders loaded on a quartz crucible (Supporting Information Figure S1). The growth process of GaN shell layers has been described in details in our previous work.52 Typically, high-purity Ar gas with a flowing rate of 200 mL/min is first used to protect the oxidation reaction before the temperature increases to 900 °C and then NH3 gas with a flowing rate of 50 mL/min is introduced into the chamber for GaN growth. The tailoring of the GaN shell thickness was realized by simply controlling the reaction time (5, 60, and 120 min, respectively) at an optimized temperature of 1100 °C. Structural Characterization and Composition Analysis. Initial 3C-SiC and WZ-GaN/3C-SiC heterostructure nanowires were examined by means of an X-ray powder diffractometer (XRD, Rigaku RINT 2000) operating at 40 kV and 40 mA by using Cu Kα radiation (λ = 1.54056 Å), a scanning electron microscope [SEM, INSPECT, F50], and a high-resolution fieldemission transmission electron microscope [(TEM) FEI, Tecnai G2 F20] equipped with an X-ray energy dispersive spectrometer (EDS). The elemental maps and line-scan profiles were recorded in the STEM mode under an accelerated voltage of 200 kV.

obvious epitaxial growth relationship with the inner 3C-SiC nanowire core (the yellow dashed line marks the interface between WZ-GaN and SiC phases). Further analysis with respect to the crystalline orientation of WZ-GaN and 3C-SiC concluded that the (0001) atomic planes (green lines) of the WZ-GaN are strictly parallel to the (111) planes (red lines) of the 3C-SiC. HRTEM imaging performed on selected areas (1 and 2) near the crystalline boundaries shown in Figure 7c indicates that WZ-GaN grains hold a given lattice matching with their neighboring 3C-SiC crystal. The atomic planar distance of 0.52 nm observed in Figure 7e,f correspond well to the d-spacing of the (0001) layers in WZ-GaN and the value of 0.25 nm is a consequence of the interplanar distance of the (111) plane of 3C-SiC. Apparently, a lattice matching between WZ-GaN and 3C-SiC with a relationship of dWZ‑GaN(0001) ≈ 2d3C‑SiC(111) has been established independently of the orientation of the GaN grains. Additionally, the orientations of the WZ-GaN grains are strictly confined by the 3C-SiC nanowires. We only observe an epitaxial growth of the (0001) plane of WZ-GaN grains on the (111) plane of 3C-SiC. Here, the included angle (70.5°) between the (0001) planes of the two WZ-GaN grains agree well with the included angle (70.5°) of two (111) planes in 3C-SiC (Figure 7c). Figure 7d shows another case of an epitaxial growth of WZ-GaN on a 3C-SiC surface (The yellow dash line marks the phase boundary of outer WZ-GaN shell and inner SiC nanowire core). As presented in Figure 2c and reported in our previous work,42 some stacking fault along the (111) plane of a 3C-SiC nanowire with a deviation angle of 19.5° along its orientation direction, occasionally forms and subsequently leads to the transition of 3C-SiC to hexagonal SiC for several atomic layers. In this case, the hexagonal SiC provides an ideal nucleation site for the epitaxial growth of WZ-GaN on its surface, and the (0001) plane of the WZ-GaN is still perfectly parallel to the (111) plane of the 3C-SiC. The measured lattice distances of 0.25 and 0.52 nm in Figure 7g correspond to the d-spacing of 3C-SiC and WZ-GaN, respectively, and match with the data shown in Figure 7e,f. All these HRTEM results obtained on the corresponding WZ-GaN/3C-SiC interfaces allow for a profound insight into the epitaxial growth mechanism of WZGaN on 3C-SiC and are able to explain the morphological evolution/wrapping process of WZ-GaN on a 3C-SiC surface. The results presented above indicate the importance of stacking faults, which are initially formed in the 3C-SiC nanowires, in guiding the preferential nucleation of GaN nanoislands and inducing their subsequent morphology evolution to a consecutive GaN shells covering on the surface of the 3C-SiC nanowires. The appearance of stacking faults will generate numerous atomic steps on the SiC nanowires surface and result in a higher surface energy in these defect sections in comparison with defect-free areas. The higher surface energy on the stacking fault section of 3C-SiC nanowires, which originates from the disorder in the arrangement of atoms, directly induces the first preferential nucleation of GaN nanoislands and realizes the shielding of GaN shells on 3CSiC nanowires to form GaN/SiC core−shell heterostructures. Such WZ-GaN/3C-SiC heterostructures with controllable morphology (GaN shells) and tunable interface provide a plethora of opportunities for their applications in the fields of nanoscaled high-speed field-effect transistor (FET), efficient photocatalysts for water-splitting to produce hydrogen, and ultrasensitive photodetector and related optoelectronic nanodevices. H

DOI: 10.1021/acs.nanolett.5b02454 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

■

(15) Dick, K. A.; Kodambaka, S.; Reuter, M. C.; Deppert, K.; Samuelson, L.; Seifert, W.; Wallenberg, L. R.; Ross, F. M. Nano Lett. 2007, 7, 1817−1822. (16) Krogstrup, P.; Yamasaki, J.; Sørensen, C. B.; Johnson, E.; Wagner, J. B.; Pennington, R.; Aagesen, M.; Tanaka, N.; Nygård, J. Nano Lett. 2009, 9, 3689−3693. (17) Nilsson, H. A.; Duty, T.; Abay, S.; Wilson, C.; Wagner, J. B.; Thelander, C.; Delsing, P.; Samuelson, L. Nano Lett. 2008, 8, 872− 875. (18) Schukfeh, M. I.; Storm, K.; Mahmoud, A.; Sondergaard, R. R.; Szwajca, A.; Hansen, A.; Hinze, P.; Weimann, T.; Svensson, S. F.; Bora, A.; Dick, K. A.; Thelander, C.; Krebs, F. C.; Lugli, P.; Samuelson, L.; Tornow, M. ACS Nano 2013, 7, 4111−4118. (19) Zhao, S.; Kibria, M. G.; Wang, Q.; Nguyen, H. P. T.; Mi, Z. Nanoscale 2013, 5, 5283−5287. (20) Kibria, M. G.; Nguyen, H. P. T.; Cui, K.; Zhao, S.; Liu, D.; Guo, H.; Trudeau, M. L.; Paradis, S.; Hakima, A.-R.; Mi, Z. ACS Nano 2013, 7, 7886−7893. (21) Beeler, M.; Hille, P.; Schörmann, J.; Teubert, J.; de la Mata, M.; Arbiol, J.; Eickhoff, M.; Monroy, E. Nano Lett. 2014, 14, 1665−1673. (22) Lim, S. K.; Brewster, M.; Qian, F.; Li, Y.; Lieber, C. M.; Gradečak, S. Nano Lett. 2009, 9, 3940−3944. (23) Li, Y.; Xiang, J.; Qian, F.; Gradečak, S.; Wu, Y.; Yan, H.; Blom, D. A.; Lieber, C. M. Nano Lett. 2006, 6, 1468−1473. (24) Hu, J. Q.; Bando, Y.; Liu, Z. W.; Sekiguchi, T.; Golberg, D.; Zhan, J. H. J. Am. Chem. Soc. 2003, 125, 11306−11313. (25) Hong, Y. J.; Kim, Y.-J.; Jeon, J.-M.; Kim, M.; Choi, J. H.; Baik, C. W.; Il Kim, S.; Park, S. S.; Kim, J. M.; Yi, G.-C. Nanotechnology 2011, 22, 205602−1−6. (26) Dick, K. A.; Deppert, K.; Karlsson, L. S.; Larsson, M. W.; Seifert, W.; Wallenberg, L. R.; Samuelson, L. MRS Bull. 2007, 32, 127−133. (27) Sun, K.; Jing, Y.; Park, N.; Li, C.; Bando, Y.; Wang, D. L. J. Am. Chem. Soc. 2010, 132, 15465−15467. (28) Day, R. W.; Mankin, M. N.; Gao, R.; No, Y.-S.; Kim, S.-K.; Bell, D. C.; Park, H.-G.; Lieber, C. M. Nat. Nanotechnol. 2015, 10, 345− 352. (29) Yang, B.; Yuan, F.; Liu, Q.; Huang, N.; Qiu, J.; Staedler, T.; Liu, B.; Jiang, X. ACS Appl. Mater. Interfaces 2015, 7, 2790−2796. (30) Meng, F.; Estruga, M.; Forticaux, A.; Morin, S. A.; Wu, Q.; Hu, Z.; Jin, S. ACS Nano 2013, 7, 11369−11378. (31) Morin, S. A.; Bierman, M. J.; Tong, J.; Jin, S. Science 2010, 328, 476−480. (32) Lau, Y. K. A.; Chernak, D. J.; Bierman, M. J.; Jin, S. J. Am. Chem. Soc. 2009, 131, 16461−16471. (33) Zhao, D. G.; Xu, S. J.; Xie, M. H.; Tong, S. Y.; Yang, H. Appl. Phys. Lett. 2003, 83, 677−679. (34) Davydov, V. Y.; Averkiev, N. S.; Goncharuk, I. N.; Nelson, D. K.; Nikitina, I. P.; Polkovnikov, A. S.; Smirnov, A. N.; Jacobsen, M. A.; Semchinova, O. K. J. Appl. Phys. 1997, 82, 5097−5102. (35) Wu, J.; Yaguchi, H.; Nagasawa, H.; Yamaguchi, Y.; Onabe, K.; Shiraki, Y.; Ito, R. Japanese Journal of Applied Physics Part 1-Regular Papers Short Notes & Review Papers 1997, 36, 4241−4245. (36) Yaguchi, H.; Wu, J.; Zhang, B. P.; Segawa, Y.; Nagasawa, H.; Onabe, K.; Shiraki, Y. J. Cryst. Growth 1998, 195, 323−327. (37) Tang, C.; Bando, Y. Appl. Phys. Lett. 2003, 83, 659−661. (38) Liu, B.; Yuan, F.; Dierre, B.; Sekiguchi, T.; Zhang, S.; Xu, Y.; Jiang, X. ACS Appl. Mater. Interfaces 2014, 6, 14159−14166. (39) Yuan, F.; Liu, B.; Wang, Z.; Yang, B.; Yin, Y.; Dierre, B.; Sekiguchi, T.; Zhang, G.; Jiang, X. ACS Appl. Mater. Interfaces 2013, 5, 12066−12072. (40) Liu, B.; Wang, Z.; Yuan, F.; Benjamin, D.; Sekiguchi, T.; Jiang, X. RSC Adv. 2013, 3, 22914−22917. (41) Shen, G. Z.; Bando, Y.; Ye, C. H.; Liu, B. D.; Golberg, D. Nanotechnology 2006, 17, 3468−3472. (42) Liu, B.; Bando, Y.; Tang, C.; Mitome, M.; Yamaura, K.; Takayama-Muromachi, E.; Golberg, D. J. Phys. Chem. C 2008, 112, 18911−18915. (43) Feng, D. H.; Jia, T. Q.; Li, X.; Xu, Z. Z.; Chen, J.; Deng, S. Z.; Wu, Z. S.; Xu, N. S. Solid State Commun. 2003, 128, 295−297.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.5b02454. Schematic diagram for describing the synthetic routines of 3C-SiC nanowires and WZ-GaN/3C-SiC heterostructures; SEM image of 3C-SiC nanowires; XRD pattern of 3C-SiC with polymorphs; STEM image, spatially resolved elemental mappings, and line-scan profiles of WZ-GaN/3C-SiC heterostructures; and TEM images of WZ-GaN/3C-SiC heterostructures with continuous GaN shells. (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was partially supported by the Knowledge Innovation Program of Institute of Metal Research with Grants Y2NCA111A1 and Y3NCA111A1, and the Youth Innovation Promotion Association, Chinese Academy of Sciences (Grant Y4NC711171). B.D.L. would also like to thank the Chinese Scholarship Council (Grant 201400260067) for the support of this work. B.Y. thanks for the National Nature Science Foundation of China (Grant 51402309). T.S. also thanks for the financial support of this work from DAAD with Grant 57054770.

■

REFERENCES

(1) Gudiksen, M. S.; Lauhon, L. J.; Wang, J.; Smith, D. C.; Lieber, C. M. Nature 2002, 415, 617−620. (2) Lauhon, L. J.; Gudiksen, M. S.; Wang, C. L.; Lieber, C. M. Nature 2002, 420, 57−61. (3) Qian, F.; Li, Y.; Gradečak, S.; Wang, D.; Barrelet, C. J.; Lieber, C. M. Nano Lett. 2004, 4, 1975−1979. (4) Verheijen, M. A.; Immink, G.; de Smet, T.; Borgström, M. T.; Bakkers, E. P. A. M. J. Am. Chem. Soc. 2006, 128, 1353−1359. (5) Xiang, J.; Lu, W.; Hu, Y. J.; Wu, Y.; Yan, H.; Lieber, C. M. Nature 2006, 441, 489−493. (6) Jiang, X.; Xiong, Q.; Nam, S.; Qian, F.; Li, Y.; Lieber, C. M. Nano Lett. 2007, 7, 3214−3218. (7) Hong, Y. J.; Jeon, J. M.; Kim, M.; Jeon, S. R.; Park, K. H.; Yi, G. C. New J. Phys. 2009, 11, 125021−13. (8) Kargar, A.; Sun, K.; Jing, Y.; Choi, C.; Jeong, H.; Zhou, Y.; Madsen, K.; Naughton, P.; Jin, S.; Jung, G. Y.; Wang, D. Nano Lett. 2013, 13, 3017−3022. (9) Higginbotham, A. P.; Larsen, T. W.; Yao, J.; Yan, H.; Lieber, C. M.; Marcus, C. M.; Kuemmeth, F. Nano Lett. 2014, 14, 3582−3586. (10) Hillerich, K.; Dick, K. A.; Wen, C. Y.; Reuter, M. C.; Kodambaka, S.; Ross, F. M. Nano Lett. 2013, 13, 903−908. (11) Shen, G. Z.; Chen, D.; Lee, C. J. J. Phys. Chem. B 2006, 110 (32), 15689−15693. (12) Wu, X.; Jiang, P.; Ding, Y.; Cai, W.; Xie, S.-S.; Wang, Z. L. Adv. Mater. 2007, 19, 2319−2323. (13) Liu, W.; Wang, R. M.; Wang, N. Appl. Phys. Lett. 2010, 97, 041916−1−3. (14) Tian, W.; Zhang, C.; Zhai, T. Y.; Li, S. L.; Wang, X.; Liu, J. W.; Jie, X.; Liu, D. Q.; Liao, M. Y.; Koide, Y.; Golberg, D.; Bando, Y. Adv. Mater. 2014, 26, 3088−3093. I

DOI: 10.1021/acs.nanolett.5b02454 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters (44) Zhou, W.; Liu, X.; Zhang, Y. Appl. Phys. Lett. 2006, 89, 223124− 1−3. (45) Devanathan, R.; Weber, W. J. J. Nucl. Mater. 2000, 278, 258− 265. (46) Okumura, H.; Misawa, S.; Yoshida, S. Appl. Phys. Lett. 1991, 59, 1058−1060. (47) Paisley, M. J.; Sitar, Z.; Posthill, J. B.; Davis, R. F. J. Vac. Sci. Technol., A 1989, 7, 701−705. (48) Vomiero, A.; Ferroni, M.; Comini, E.; Faglia, G.; Sberveglieri, G. Nano Lett. 2007, 7, 3553−3558. (49) Starke, U.; Bram, C.; Steiner, P. R.; Hartner, W.; Hammer, L.; Heinz, K.; Muller, K. Appl. Surf. Sci. 1995, 89, 175−185. (50) Kimoto, T.; Itoh, A.; Matsunami, H. Appl. Phys. Lett. 1995, 66, 3645−3647. (51) Addamiano, A. J. Cryst. Growth 1982, 58, 617−622. (52) Liu, B. D.; Bando, Y.; Tang, C. C.; Xu, F. F.; Hu, J. Q.; Golberg, D. J. Phys. Chem. B 2005, 109, 17082−17085.

J

DOI: 10.1021/acs.nanolett.5b02454 Nano Lett. XXXX, XXX, XXX−XXX