Atomistic structure simulation of silicon nanocrystals driven with suboxide penalty energies

The structural control of silicon nanocrystals embedded in amorphous oxide is currently an important technological problem. In this work, an approach is presented to simulate the structural behavior of silicon nanocrystals embedded in amorphous oxide matrix based on simple valence force fields as described by Keating-type potentials. After generating an amorphous silicon-rich-oxide, its evolution towards an embedded nanocrystal is driven by the oxygen diffusion process implemented in the form of a Metropolis algorithm based on the suboxide penalty energies. However, it is observed that such an approach cannot satisfactorily reproduce the shape of annealed nanocrystals. As a remedy, the asphericity and surface-to-volume minimization constraints are imposed. With the aid of such a multilevel approach, realistic-sized silicon nanocrystals can be simulated. Prediction for the nanocrystal size at a chosen oxygen molar fraction matches reasonably well with the experimental data when the interface region is also accounted. The necessity for additional shape constraints suggests the use of more involved force fields including long-range forces as well as accommodating different chemical environments such as the double bonds.