Applied Surface Science, Vol.464, 321-327, 2019
Evolution of medium-range order and surface compositions by mechanism-driven model with realistic network
To elucidate the evolution of structural topology with the medium-range order of growing silicon films fabricated by plasma enhanced chemical vapor deposition with highly diluted hydrogen, we established a the model for a realistic network of growth films based on continuous random networks. We combined the kinetic Monte Carlo method with the interactions of various atomic-scale mechanisms from first-principles density-functionaltheory computations and molecular-dynamics computations. To quantitatively characterize the short- and medium-range order, in addition to the higher order of the structural network, we applied a fluctuation transmission electron microscopy simulation and the pair correlation function to measure information about local order regions in film network. Interestingly, we found that the inflexion temperature of surface SiH3 coverage directly affected the silicon hydride crystallization process by forming much chemisorption from physisorption of surface hydride species. More interestingly, based on the atomic-scale chemical mechanism of the growth process of film, our results first realistically rendered the continuous disorder-order phase transition from structural topology, which is differed from previous knowledge that the amorphous to polycrystalline transition is a discontinuous short-range-order to long-range-order phase transition. Especially, we predicted the temperature dependence of evolution of structural network of the film and elucidated that the growing silicon film formed a nanocrystalline structural network at very low temperatures, relative to those of thermal annealing. This result was ascribed to interactions and competitions between various mechanisms and critically due to the dissociation mechanism and H-induced crystalline mechanism of silicon hydride, which isn't depend on only one of the model mechanisms of surface diffusion model, etching model and annealing model. These results provide significant new physical insight into an experimentally relevant process for optimizing a strategy of deposition condition.