The early growth stages of distinctive physical vapor deposition (PVD) techniques [namely molecular beam epitaxy (MBE), sputtering, pulsed thermal evaporation -e.g., flash evaporation (FE) and pulsed laser deposition (PLD)] are investigated by rate equation (RE) approach.[1] The surfaces that result from these techniques are studied in regard to: their roughnesses, number of uncovered layers forming each surface and the distribution per layer of the average size of islands that act as nucleation surface for upper layers. For such an analysis, the RE approach was improved to account for (besides standard mechanisms [1,2]) the following processes: i) the coalescence of islands, ii) the island-size dependence of the step-edge barriers, and iii) the nucleation of successive layers considering the confined diffusion on the islands.
The solutions of the RE system provide the following results: The pulsed growth techniques (FE and PLD) produce flatter surfaces than those obtained from continuous fluxes (MBE and sputtering). The lower roughnesses of the former result from an increase in the density of stable moderate-sized clusters, which are small enough to prevent the nucleation of upper layers prior to layer completion. An increase in the kinetic energy of the incident species within the pulsed flux induces a further decrease in the surface roughness, which explains why PLD is par excellence PVD technique to grow the flattest films.[3] Conversely, continuous fluxes of energetic species (sputtering) produce an extra increase in the roughness with respect to the MBE surface. The simulations show that the origin of this divergent behavior resides in a detailed balance between two complementary mechanisms: i) the island coarsening by coalescence of smaller species (that prevails for continuous fluxes giving rise to a Poisson cluster-size distribution) and ii) the coarsening by Ostwald ripening (typical of pulsed growth).
These results play a key role for the surface engieneering and thin film nanostructuration revealing features and inherent limitations in some of the mostly used techniques for materials growth.
1) E. Vasco, New J. Phys. 8, 253 (2006)
2) E. Vasco et al., Phys. Rev. Lett. 98, 036104 (2007)
3) E. Vasco et al., J. Phys.-Cond. Mat. 16, 8201 (2004) |