Anisotropy in Nucleation and Growth of Two-Dimensional Islands during Homoepitaxy on "Hex" Reconstructed Au(100)

Abstract

We present results of a comprehensive scanning tunneling microscopy study of the nucleation and growth of Au islands on Au(100). It is shown that the reconstruction of the substrate produces strong anisotropic effects. Rate equation analysis of the experimental flux and temperature dependence of the island density suggests: (i) a critical size of i = 3 for T = 315 -380 K, but i ) 3 above 400 K; and (ii) strongly anisotropic diffusion, preferentially parallel to the reconstruction rows (activation energy -0.2 eV). We comment on energetic and kinetic aspects of the observed island shape anisotropy. PACS numbers: 61.16.Ch, 68.55.a Nucleation and growth of islands during thin film deposition has been studied for decades. Traditional analyses, using mainly scanning and transmission electron microscopy, concentrated on systems where three-dimensional islands form [1]. The results were generally interpreted via rate equations. Recent revival of interest is associated with the availability of scanning tunneling microscopy (STM) [2 -7] and high resolution surface sensitive diffraction techniques [8 -11],which are especially suited for analyses of nucleation and growth of two-dimensional (2D) islands on atomic length scales.Here we provide results of the first comprehensive study of anisotropic nucleation and growth in a metalon-metal system. STM observations on the homoepitaxial growth on the reconstructed Au(100) surface at temperatures between 31S and 43S K are analyzed quantitatively in a rate equation approach. A key component of our analysis is the demonstration that diffusion is actually strongly anisotropic for Au on Au(100). This is a consequence of the reduced symmetry of the reconstructed Au(100) surface. To this end, we must extend the traditional mean-field rate equations [1] to treat strongly anisotropic diffusion. Comparison with "exact" Monte Carlo simulation results is also used. In this system, we observe noticeable denuded zones at steps orthogonal, but not parallel, to the reconstruction rows. Analogous behavior for Si/Si(100) homoepitaxy was attributed to strongly anisotropic diffusion [2]. In our analysis we do not preclude a priori the possibility that the observed anisotropy is caused by different sticking coefficients (anisotropic bonding) at differently oriented step edges. Instead we examine data for the dependence of the average density of islands N, on the deposition flux R and substrate temperature T at fixed submonolayer coverages, thereby deducing diffusion anisotropy, the probable values of the critical island size i, at and above room temperature, as well as estimates of the binding energy of critical islands and the energy barrier for isolated adatom diffusion.The experiments were performed in a UHV system with standard facilities for sample preparation and characterization and a pocket-size STM (base pressure 6 x 10 9 Pa).

Electronic structure of reconstructed Au(100): Two-dimensional and one-dimensional surface states

“…In Ref. 41, Au islands grown on reconstructed Au(100) tend to show widths of 30Å or less, and lengths range from 100 to 150Å, figures that fit well with the reconstructed unit cell. Thus, five atom rows ×2.88Å = 14.4Å, so two reconstruction rows ≈30Å.…”

“…[36][37][38][39][40] This is surely due to a charge transfer from the surface into the adsorbed species that reverses the mechanism, electronic in origin, that triggers reconstruction. On the other hand, there are experimental 41 and theoretical 42 evidences that point that Au(100) reconstructed areas have a "magic width," a multiple of five substrate atom rows (14.4Å), the width of the reconstruction fringes of Fig. 1(a).…”

Models for adatom diffusion on fcc (001) metal surfaces

“…In some experiments STM was used to acquire information on larger scales, e.g., island size distributions [12,14,[20][21][22]26]. Despite the wealth of experimental results at the atomic scale, for decay rates of small islands, mobility of small islands and edge diffusion [21,[23][24][25][26][27]29,31], the underlying energetics is mostly inaccessible to direct experimental measurements. Thus, one must rely on theory to extract activation energies from the experimental results, and these are usually limited in number, and sometimes are subject to alternative interpretations.…”

Initial stages of Cu epitaxy on Ni(100): Postnucleation and a well-defined transition in critical island size

“…1,2 The saturation island density can be measured by several surface-sensitive techniques, such as electron microscopy, 2 electron diffraction, 3 helium atom diffraction, 4 and scanning tunneling microscopy ͑STM͒. [5][6][7][8][9][10][11][12] With respect to STM, the other methods have certain disadvantages. The resolution of electron microscopy is not sufficient to image high island densities (n x Ͼ10 Ϫ3 ͒ and small islands of a few atoms.…”

“…8 The analysis of nucleation on anisotropic substrates, however, is still in progress. 9,10,12 The particular bond geometry on square lattices leads to characteristic features in the nucleation kinetics on fcc͑100͒ surfaces. These particularities include the stability of islands.…”