An interpretation of the temperature-dependent growth mode of copper on the Cu(001)(2√2×√2)R45°-O surface

Abstract

The growth mode of copper during homoepitaxial growth on the Cu(001)(2√2×√2)R45°-O surface has been discussed on the basis of a phenomenological (a simple statistical thermodynamic) treatment of the order–disorder arrangement of the missing-row structure. In this treatment the oxygen surface coverage, θ, is fixed at 0.5 ML (monolayer), and the surface free energy is calculated as a function of the concentration of copper vacancies on the first surface layer. Based on the fact that the missing-row structure, in which every fourth topmost copper row is deficient, is the most stable structure thermodynamically, the surface free energy is expressed as a function of copper vacancies. According this treatment, it is shown that only 1/4 ML copper deposition is sufficient to lose the Cu(001)(2√2×√2)R45°-O reconstruction structure provided that impinging copper atoms predominantly occupy missing-row copper sites. This is compatible with the disappearance of (1/2,1/2) and (3/4,3/4) reflective high-energy electron diffraction superlattice reflections at low substrate temperatures and/or at high deposition rate. At higher substrate temperature (>510 K), the reaction proceeds to keep the surface free energy a minimum; i.e., to keep the (2√2×√2)R45°-O surface structure of one step height difference. Consequently, the growth mode becomes step-flow growth or layer-by-layer growth depending on the temperature and deposition rate. In this step flow or layer-by-layer growth, the assumption of different copper capture probabilities on each row, e.g., 1/2 for a normal row, 1 for missing a row, 0 or 1/2 on an oxygen-adsorbed row, is essential.