Effects of Crystalline Microstructure on Epitaxial Morphology
- PDF / 1,335,133 Bytes
- 10 Pages / 414.72 x 648 pts Page_size
- 81 Downloads / 313 Views
the existing models is the unphysical assumption that growth takes place on a regular (generally simple-cubic) lattice with the solid-on-solid (SOS) growth rule in which atoms move only along lattice directions and land on top of one another without taking the crystal structure of the substrate or the growing material into account. These unphysical features lead to a number of pathological morphologies [121 and dynamics which do not agree with experiments. For example, because of the epitaxial nature of the growth there exists a crucial dynamical process which initially must take place when an atom randomly lands onto the surface. This is the process by which some of the initial energy of condensation of the deposited atom is dissipated as it moves and relaxes through a series of cascades until it reaches an epitaxial site on the surface [13]. Such a microscopic process is not naturally built into lattice-SOS models, but it fundamentally alters the growth and the morphology of the surface by limiting the steepest angles on the surface. These simple cascade processes in turn eliminate the problems of large cliffs, infinite vertical diffusion and angle selection that are inherent to the SOS models. Here we present the results of kinetic Monte Carlo simulations of epitaxial growth that take the crystal structure correctly into account [14]. We find that cascades and angle selection are naturally built into the model and the surface morphology does not develop the sort of unphysical groove instability that is well known in studies of SOS models [3], [4]. We may apply our technique to model epitaxial growth in metal bcc(100) and fcc(100) systems. In particular we have used our model to study multilayer growth in Fe/Fe(100) at room temperature, which has been the subject of several recent experiments [15], [16]. We find that the existence of the Ehrlich-Schwoebel barrier [17] to interlayer diffusion leads to the formation and coarsening of mounds as observed in experiments. Our results are in excellent agreement with the development and the value of the selected mound angle as observed in Scanning Tunnelling Microscope (STM) images of the surface morphology [16]. We also find excellent agreement with the experimental results [15], [16] for the coarsening and roughening of surface features. MODEL AND SIMULATIONS In our model atoms are randomly deposited at a rate F/2 per lattice site per unit time (corresponding to F layers per unit time) onto a square lattice (see Fig. l(a) ) corresponding to the bcc(100) substrate used in our simulations. Due to the epitaxial nature of the growth, deposited atoms are incorporated into the system only at the four-fold hollow sites formed by the four nearest-neighbor atoms in the layer below. This implies that if the deposited atom lands directly on a four-fold hollow site (see Fig. l(c) ) then it becomes part of the surface. However, if one or more of the nearest-neighbor sites has height lower than one less than the height of the deposited atom ( Fig. l(b) ) then the freshly deposited
Data Loading...
