A morphological study of vicinal surface growth based on a hybrid algorithm

A hybrid algorithm that combines a phase-field model and a lattice gas model evolving according to a kinetic Monte-Carlo (KMC) simulation scheme is used to investigate the dynamics of vicinal surface growth during vapor phase epitaxy. The algorithm is computationally far more efficient than pure KMC schemes, and this gain in efficiency does not correspond to a loss in information on the kinetics of individual atoms. We present numerical studies on the temperature dependence of macroscopic properties of the growing surface, evaluating the relevant stochastic processes (attachment, detachment, diffusion and island dynamics) as a function of their rates. We show that the temperature at which step flow is replaced by island nucleation depends on incoming flux, diffusion parameters and interstep distance. Moreover, we validate these finding by comparison to experiments and by analytical investigations.