We develop a nucleation-based model to explain the formation of the wurtzite (WZ) phase during the vapor-liquid-solid growth of free-standing nanowires of zinc-blende (ZB) semiconductors. Nucleation occurs preferentially at the edge of the solid/liquid interface, which entails major differences between ZB and WZ nuclei. Depending on the pertinent interface energies, WZ nucleation is favored at high liquid supersaturation. This explains our systematic observation of ZB during early growth.PACS numbers: 68.65. La,64.60.Qb,81.05.Ea,81.15.Kk,64.70.Nd Free-standing wires with diameters ranging from hundreds down to a few nanometers are nowadays commonly fabricated from a large range of semiconductor materials [1,2,3,4,5]. These nanowires (NWs) have remarkable physical properties and many potential applications. The present work deals with the epitaxial growth of NWs of III-V semiconductors on a hot substrate. Metal catalyst nanoparticles deposited on the substrate before growth define the wire diameter. According to the vapor-liquidsolid (VLS) growth mechanism, the atoms are fed from the vapor phase to the solid wire through this particle (or droplet), which remains liquid during growth [6].We consider III-V compounds which, under bulk form, adopt the cubic zinc-blende (ZB) crystal structure [7] (although some non-ZB high-pressure phases [8] may be metastable at atmospheric pressure [9]), leaving aside nitrogen-based NWs. We discuss the usual case of NWs grown on a [111]B (As-terminated) face of the ZB substrate. Probably the most surprising feature of these NWs is that, in contrast to their bulk counterparts, they often adopt the hexagonal wurtzite (WZ) structure. This was observed for most ZB III-V materials and growth techniques [1,3,4,10,11]. However, although often dominantly of WZ structure, the NWs usually contain stacking faults (SFs) and sequences of ZB structure. The coexistence of two phases is clearly a problem for basic studies as well as applications, so that phase purity control is one of the main challenges of III-V NW fabrication.The surprising prevalence of the WZ structure in III-V NWs has not been explained satisfactorily so far. Here, based on new experimental observations, we propose an explanation of the occurrence of the WZ structure and develop a model predicting quantitatively in which growth conditions it should form. We consider the specific case of gold-catalyzed GaAs NWs grown by molecular beam epitaxy (MBE) on a GaAs substrate but we expect our model and our conclusions to remain valid for any ZB III-V compound and any growth method.Let us start with briefly reviewing previously proposed explanations. Calculations give the difference δw in cohesive energy between ZB and WZ bulk GaAs as about 24 meV per III-V pair at zero pressure [7]. It has been argued that this favoring of the ZB form might be offset in NWs of small diameter by the large relative contribution to the total energy of either the lateral facets [12] or the vertical edges separating the latter [13] (provided the specifi...

No abstract

No abstract

In semiconducting nanowires, both zinc blende and wurtzite crystal structures can coexist. The band structure difference between the two structures can lead to charge confinement. Here we fabricate and study single quantum dot devices defined solely by crystal phase in a chemically homogeneous nanowire and observe single photon generation. More generally, our results show that this type of carrier confinement represents a novel degree of freedom in device design at the nanoscale.

GaAs nanowires were grown by molecular-beam epitaxy on (111)B oriented surfaces, after the deposition of Au nanoparticles. Different growth durations and different growth terminations were tested. After the growth of the nanowires, the structure and the composition of the metallic particles were analyzed by transmission electron microscopy and energy dispersive x-ray spectroscopy. We identified three different metallic compounds: the hexagonal β′Au7Ga2 structure, the orthorhombic AuGa structure, and an almost pure Au face centered cubic structure. We explain how these different solid phases are related to the growth history of the samples. It is concluded that during the wire growth, the metallic particles are liquid, in agreement with the generally accepted vapor-liquid-solid mechanism. In addition, the analysis of the wire morphology indicates that Ga adatoms migrate along the wire sidewalls with a mean length of about 3μm.

We report on the new mode of the vapor-liquid-solid nanowire growth with a droplet wetting the sidewalls and surrounding the nanowire rather than resting on its top. It is shown theoretically that such an unusual configuration happens when the growth is catalyzed by a lower surface energy metal. A model of a nonspherical elongated droplet shape in the wetting case is developed. Theoretical predictions are compared to the experimental data on the Ga-catalyzed growth of GaAs nanowires by molecular beam epitaxy. In particular, it is demonstrated that the experimentally observed droplet shape is indeed nonspherical. The new VLS mode has a major impact on the crystal structure of GaAs nanowires, helping to avoid the uncontrolled zinc blende-wurtzite polytylism under optimized growth conditions. Since the triple phase line nucleation is suppressed on surface energetic grounds, all nanowires acquire pure zinc blende phase along the entire length, as demonstrated by the structural studies of our GaAs nanowires.

Self-catalyzed growth of GaAs nanowires by molecular beam epitaxy on (111)Si substrates is investigated by introducing Al x Ga 1−x As time markers. The nanowire elongation rate is found to be radius-independent, constant at substrate temperatures below 650°C and linearly increasing with the incoming arsenic flux. The basic question of which pathways are followed by the arsenic species contributing to nanowire growth is clarified. The flow rate of As atoms directly impinging on the Ga catalyst drop is significantly smaller than the As consumption by nanowire growth. Thus, supplementary As atoms are necessary to explain the actual elongation rate. We show that surface diffusion of adsorbed As x species toward the catalyst cannot account for the missing atoms. On the other hand, the reevaporation of As x species from the substrate and from nanowire sidewall surfaces can act as an efficient secondary arsenic source. We argue that a sufficient amount of these species can be intercepted by the Ga drop and add up with the direct As impingement to explain the actual elongation rate.

The effect of sidewall nucleation on nanowire morphology is studied theoretically. The model provides a semiquantitative description of nanowire radius as a function of its length and the distance from the surface. It is demonstrated that the wire shape critically depends on the diffusion flux of adatoms from the substrate and on the rate of direct impingement to the sidewalls. At high diffusion flux the wire shape is cylindrical. A decrease of diffusion from the surface leads to the onset of nucleation on the sidewalls resulting in the lateral extension and in the reduction of wire length. The wire shape changes from cylindrical to conical, because the supersaturation of adatoms driving the nucleation is higher at the wire foot than at the top. It is shown that the shape modification becomes pronounced at low growth temperatures. Theoretical results are used to model the experimentally observed shapes of GaAs and GaP wires, grown by Au-assisted molecular beam epitaxy at different temperatures.