Gold nanoclusters on silicon dioxide are interrogated by charge-state resolved Low Energy Ion Scattering (LEIS). The LEIS data are analyzed in a novel way that quantitatively reveals information about the homogeneity of the clusters’ size distribution. The method utilizes the fact that the neutralization probability and the probability of multiple scattering (MS) both depend on the cluster size, but in different ways. Smaller nanoclusters are more likely to neutralize scattered alkali ions [1], while larger nanoclusters lead to more MS. Thus, in a system with a bimodal distribution of cluster sizes, multiply scattered ions are less likely to be neutralized. This leads to a difference in the spectral shapes of the scattered ions and neutrals. In contrast, a system with a homogeneous cluster size distribution yields identical spectral shapes. LEIS has advantages over the use of STM for measuring cluster size distributions, as it is easier to deploy over a wide range of preparation conditions and STM cannot probe disordered materials.

We collected time-of-flight LEIS spectra for 2 keV Na⁺ ions scattered from Au nanoclusters formed by deposition onto SiO₂ at room temperature and after annealing to increasing temperatures. The inhomogeneity in the cluster size distribution is quantified with a metric obtained by analyzing spectra collected for different charge states of the scattered projectiles. It is found that the cluster sizes are fairly uniform after the initial deposition, but a heterogeneous distribution, presumably due to Ostwald ripening, develops after annealing above 700 K.

[1] G. F. Liu, Z. Sroubek, and J. A. Yarmoff, Phys. Rev. Lett. 92, 216801 (2004).

In submonolayer epitaxial island growth, it is fruitful to consider the distribution of the area of capture zones, i.e. Voronoi (proximity) cells constructed from the island centers. For random nucleation centers (Poisson Voronoi diagrams) the CZD is expected to follow a Gamma distribution, but more generally we have argued [1], drawing from experiences analyzing the terrace-width distributions of vicinal surfaces, that the CZD is better described by the single-parameter generalized Wigner distribution (GWD). Painstaking simulations by Amar's and Evans's groups showed inadequacies in our mean field Fokker-Planck argument relating the characteristic GWD exponent β to the critical nucleus size (conventionally called i+1), i.e. the size of the smallest cluster assumed not to decay. We refine our derivation to retrieve their finding that β is nearly i + 2 [2]. While the GWD describes the CZD in the regime in which there is significant data in experiments (i.e. between half and twice the mean), it has shortcomings in the tails at both high and low areas. For large areas, the distribution may decay exponentially rather than in gaussian fashion. We discuss several treatments of this issue, emphasizing the fragmentation model we developed [3], which depends on two physically motivated scaling exponents.

We discuss applications of this formula and methodology to experiments involving Ge/Si(001), various organics on SiO₂, and para-hexaphenyl (6P) films on amorphous mica. We report a series of studies by Fanfoni et al. of InAs quantum dots on GaAs and very recent applications to metallic droplets by Millunchick's group, also on GaAs. (The former also shows that the more-often-probed island-size distribution is comparable the CZD at lower temperatures but not at higher temperatures when detachment--and consequent coarsening--becomes important.)

We have also used the GWD framework to elucidate kinetic Monte Carlo studies of homoepitaxial growth on Cu(100) with codeposited impurities of different sorts. Finally, we have applied this approach to the distribution of metro stations in Paris [3] and to the distribution of of the areas of French districts (arrondissements) [3,4] , counties in southeastern US states [4], and other such secondary administrative units.

*Supported by NSF-MRSEC at UMD, Grant DMR 05-20471 and NSF-CHE Grant 07-50334.

[2] Alberto Pimpinelli and TLE, Phys. Rev. Lett. 104, 149602 (2010)

[3] Diego Luis González and TLE, Phys. Rev. E 84, 051135 (2011)

[4] Rajesh Sathiyanarayanan, Ph.D. thesis, U. Maryland, 2009; RS et al., preprint.

Silicon and germanium nano-dots self-organized on silicon surfaces are interesting from the point of view of silicon-based optoelectronic devices. In the size range from a few nm to a few tens of nanometers such structures show size-dependent electronic and optical properties [1].

Ge nano-dots have been grown on Si(111) covered by a thin oxide layer (~0.8 nm) using a wedge shaped deposition profile, resulting in varying nano-dot size along the sample profile. The prescense of the thin oxide triggers growth of nano-dots rather than flat Ge domains.

Samples were investigated by optical second harmonic generation (SHG) and photoemission spectroscopy (PES). Characterization of the growth of nano-dots with core level PES showed the decaying Si2p signal from the substrate and the increasing Ge3d signal from the growing amount of deposited Ge along the sample. The core level spectra confirm the growth mode discussed in the literature where it is suggested that contact is created between the Ge nano-dots and the Si substrate through the tin oxide [1]. Scanning electron microscopy on focused ion beam cut cross sections of the sample show that the nano-dots are largely spherical. Valence band PES shows that the position of the valence band maximum depends on nano-dot size in agreement with previous results in the literature [1].

Investigations of this system with SHG show that this technique provides a coherent surface sensitive optical probe of Ge nano-crystals on a Si surface. The signal from the nano-crystals is clearly separated from that of the substrate. It is found that the Si substrate resonance grows with Ge deposition, probably due to charges causing field induced SHG. The resonance in the Ge SHG signal can be ascribed to interband transitions at the L-point of the Ge band structure, usually referred to as the E1 critical point. With increasing particle size the Ge SHG resonance shifts towards lower energy and approaches the position expected for a plane surface. It should thus be noted that while size effects in valence band PES originate from states close to the Fermi level at the Γ- point the initial states for the SHG resonance are ~1.8 eV further down in energy at the L- point.

[1] A. A. Shklyaev and M. Ichikawa, Surface Science 514, 19 (2002).

*Partially supported by Conacyt-Mexico.

Self-organized nanostructured surfaces have been the object of much interest in recent years due to their potential applications as possible alternative to standard lithography techniques. Kern et al. [1] demonstrate that a submonolayer coverage of oxygen deposited on Cu(110) may form periodic stripes aligned along the [001] direction and that the periodicity of the system depends on the coverage. However, although this method appears very promising for the growth of nanostructured materials, the practical applications are limited due to the relatively small domains of periodicity and size as defined by the Marchenko-Vanderbilt (MV) model [2]. For oxygen coverage θ_O ranging from 0.1 < θ_O < 0.4 (the full coverage corresponds to θ_O = 0.5), the periodicity only varies from 6.5 nm to 11 nm and the oxide band width from 2 nm to 9 nm. We will show that these ranges can be significantly increased by the co-adsorption of small amounts of sulphur at the surface.

The new preparation method presented here, which consists in the partial coverage of the surface by sulphur before annealing of the oxidized surface, enables us to significantly enlarge the domain of accessible periodic structures. Periodicity up to 200 nm and oxide band width up to 30 nm can be reached in a straightforward, reproducible and controlled way. The band width and periodicity dispersions are not modified by the sulphur adsorption. Moreover, while the structure of the O/Cu(110)-(2×1) is fully determined by the surface coverage, our method allows us to define the periodicity independently of the oxygen coverage by adjusting the sulphur amount at the surface. A new model has been proposed in order to take into account the effect of the sulphur on the elastic constant of the system. This model allows a reinterpretation of results from the literature [3].

This new system may be used as a template for the growth of nanostructures as well as more fundamental purposes. Indeed, the use of such a versatile structure should enable to gain information e.g. on the elastic and electrostatic properties that are responsible for the nanostructuration of the surface or on the effect of quantum confinement by quasi-one-dimensional structure and how it is influenced by structural modification at the nanometer scale. Moreover, it may be seen as an ideal playground to test the change in the reactivity of a surface as a function of its structure.

[1] K. Kern, H. Niehus, A. Schatz, P. Zeppenfeld, J. Goerge, G. Comsa, Phys. Rev. Lett. 67, 855 (1991).

[2] V. I. Marchenko, JETP Lett 55, 73 (1992) ; D. Vanderbilt, Surf. Sci. 268, L300 (1992).

[3] K Bobrov, L. Guillemot, Surf. Sci. 604, 1894 (2010).

Stable metal nanoparticles grown on metal-supported graphene can in some circumstances form a periodic array and have potential applications in fields such as catalysis and nano-electronics. To understand their formation mechanism, we studied the adsorption and diffusion of monomers, dimers, and trimers of Rh and Au on graphene/Ru(0001) using Density Functional Theory (DFT). These two metal species exhibit distinct behaviors in cluster formation on graphene/Ru(0001). We determined the global minimum structure of Au₈ cluster on graphene/Ru(0001) using genetic algorithm. Our results on adsorption and diffusion of small metal clusters M_n (n=1, 2, 3) and Au₈ have given us insights into the nucleation process of metal clusters on graphene moiré and aid us on building a general model to describe the cluster growth mechanism on graphene moiré, which will provide guidance on designing supported sintering-resistant nanoclusters.