Computational Studies of Lattice Gas Models

We have studied the effect of diffusion on the dynamics of desorption on a lattice-gas model The chemical potential during the adsorption phase was varied in order to vary the initial size distribution for the desorption phase. The effect of diffusion on the size distribution dynamics during the desorption process was observed by turning the diffusion on and off while keeping the chemical potential during desorption constant. We study the effect of diffusion on correlation length during the desorption process. We also compare size distribution dynamics without and with the application of diffusion and study the effect of diffusion on the size distributions at given coverages. During the desorption process, the correlation length increased up to a maximum and then decreased. We found that diffusion tends to increase correlation length at any given coverage. However diffusion increase correlation length by very small percentage in the regime where correlation length is decreasing, and increase it more significantly when the correlation length is increasing. When the correlation length of the initial configuration is large, the correlation length during the desorption only increase slightly at initial coverages and decrease for the most part. As a result, diffusion only increase correlation length insignificantly during the whole process. When the correlation length of the initial configuration is small, the correlation length increase during a significant part of the process. As a result, diffusion increase correlation length significantly during the process. By studying the size distributions at some coverages during the process -before and after diffusion- we found that diffusion tends to shrink large clusters and grow or create small clusters. When the clusters growth or creating of new clusters by diffusion is small, the increase of correlation length by diffusion is small and large otherwise. We also study Bromine and Chlorine chemisorption on a Ag(100) surface, using a lattice-gas model and the quantum-mechanical Density Functional Theory (DFT) method. In this model the Br and Cl ions adsorb at the fourfold hollow sites of the Ag(100) surface, which can be represented by a square lattice of adsorption sites. Five different coverages were used for each kind of adsorbate. For each adsorbate and coverage, we obtained the minimum-energy configuration, its energy, and its charge distribution. From these data we calculated dipole moments, lateral interaction energies, and binding energies. Our results show that for Br the lateral interactions obtained by fitting to the adsorption energies obtained from the DFT calculation are consistent with long-range dipole-dipole lateral interactions obtained using the dipole moments calculated from the DFT charge distribution. For Cl we found that, while the long-range dipole-dipole lateral interactions are important, short-range attractive interactions are also present. Our results are overall consistent with parameter estimates previously obtained by fitting room-temperature Monte Carlo simulations to electrochemical adsorption isotherms [I. Abou Hamad et al., J. Electroanal. Chem. 554 (2003), 211; Electrochim. Acta 50 (2005), 5518].