Lateral ordered surface alloy

In addition to the strong reconstructions of the A1 substrate in these single and multi-layer surface alloys, further weak reconstructions are found, in which A1 atoms in the first few layers are subject to both vertical and lateral displacements from their bulk positions. Accurate treatment of these weak reconstructions is important in both LEED and DFT analyses in order to achieve agreement with experiment. Thus the alkali on aluminum system exhibits a wide spectrum of adsorbate-induced reconstructions of the substrate. 

A very similar transformation of the original hep adlayer to a surface alloy coverage with the same Tl-Tl interatomic distances and [V3 x V3]R30° symmetry has been observed in the system Tl/Ag(lll) during extended polarization of the incompletely formed first T1 adsorbate layer. As in the system Pb/Ag(lll), there is strong evidence that the transformations proceed from the boundaries of the peripheral adsorbate-free domains inwards on the terraces. However, in contrast to the system Pb/Ag(lll), the transformed coverages include both ordered and disordered domains, and then-desorption results in the formation of monoatomic pits in the substrate with widths of ca. 3 to 10 nm [3]. These pits diminish and finally vanish within a few minutes by coalescence and lateral displacement, at a rate that can be increased markedly by positive shift of the substrate potential. 

Naked tin adatoms were also obtained on a Pt/Si02 catalyst but, in order to remove all the butyl Ugands the grafted organotin complex was further heated at 300 °C. In agreement with the gas evolution of 4 butane equivalents per tin, EXAFS data indicated that tin was only surrounded by four platinum atoms at the same distance of 0.276 nm (Fig. 18.7). This result clearly indicated that tin is located on the metal surface and not in the bulk for example, in a bulk PtsSn alloy, tin is surrounded by 12 platinum atoms, while in a surface alloy on a platinum bulk it is surrounded by 6 platinum atoms only. Note that such tin adatoms are always obtained when low amounts of tin are deposited on the metal. This is probably because tetrabutyl tin coordinates first on the metal atoms of the faces which are the most hydrogenolyzing. This fact will be very important in catalysis since it explains selective poisoning of metal particles (see later). 

The last section will concentrate on the lateral distribution of the respective metals in a surface alloy. We will exemplarily show how the atom distribution in a disordered surface alloy can be quantitatively characterized based on scanning tunnehng microscopic (STM) data and how such a distribution can be predicted by Monte Carlo (MC) simulations. This will include the description of a simplified pairwise interaction model and how the energy parameters for such a model can be derived from both experiments and ab initio calculations. We will show that even a very basic energy model is capable of accurately predicting the atom distribution in a surface alloy via the MC simulations. The MC simulations also allow prediction of the (hypothetic) surface structure at temperatures where sluggish kinetics suppresses reorganization of the atoms in an experiment A key parameter to be derived from such simulations is the temperature of the order-disorder transition of the respective system. 

NNs over the entire range of film compositions investigated. This preference is partially reversed for a distance of 2 NN, where a small local maximum is visible for the surface with 25% Pd, indicating a preference for Pd atoms separated by 2 NN units, as they are present in the (2 x 2) structure on the (111) plane of ordered CujPd bulk alloys. In an SRO analysis, such a (2 x 2) structure would generate values of a(2) = l and 0 (r) = —1/3 for r 2. In the STM-based data, however, the variations of a(r) are much less enhanced, and for r>2, they lie within the margins of the statistical uncertainty. Even prolonged anneahng at lower T was shown to not yield any periodic superstmcture [57]. Chapter 11 will include a thermodynamic description of the lateral atom distribution, and it wiU show why ordered structures are experimentally inaccessible to most surface alloys. 

Using the Vienna ah initio simulation package (VASP) [93], one can calculate the total energies of ordered Au Ptj /Pt(lll) slabs. In the case shown here, 41 ordered Au Ptj /Pt(111) structures were used as input [50]. The surface alloys are represented by Au, Pt, or Au Ptj. monolayers on top of five-layer Pt(lll) slabs. The overlayers as well as the two underlying Pt layers are allowed to relax vertically and laterally. Depending on the Au/Pt ratio, the unit cells of the periodically repeated slabs had the lateral dimensions (4 x 4), ( 7 x 7), (3 x 3), (2 x 2), ( 3 x i), or (2 X 1). A few selected examples are depicted in Figure 12.27, along with the... 

Samples are taken from the melt and analyzed, and the chemical composition of the steel can, if necessary, be modified by the addition of alloying metals in the converter or in the ladle afterwards. Later, the desired molten metal is either cast into ingots or continually cast into a slab or billet form. Then the material is hot-rolled or forged into its final form. Some material receives cold rolling to further reduce the thickness as in sheets or drawn into smaller diameters as in rods and wire. Most stainless steels receive a final annealing and acid pickling in order to remove furnace scale from annealing, and they help to promote the passive surface film that naturally occurs. 

---