Fluctuation electron microscopy of medium range order in ion-implanted amorphous silicon.

摘要

Fluctuation electron microscopy is a very practical technique for the study of amorphous structure. This technique is sensitive to high-order atomic correlations. Basically, it measures medium-range order through the variance of dark-field images. In this study, fluctuation microscopy with dark-field illumination is used to probe medium-range order in ion-implanted amorphous silicon.; Ion implantation, an important precursor to semiconductor doping, can produce amorphous silicon. We find that this material can have a high degree of medium-range order. This order corresponds to a paracrystalline structure, i.e. crystallites with a size of 1–3 nm are topologically connected to a disordered structure. It is known that the as-implanted state is very unstable. The paracrystalline state is a high-energy state because it can relax towards the ideal random state on thermal annealing. The randomization is correlated with thermal relaxation seen in calorimetry, and this conclusion is furthermore supported by the similar kinetics between these two processes. In the study of post-anneal implantation, medium-range order can be reproduced by implanting heavy, energetic ions in annealed silicon. This finding resolves an apparent contradiction about the as-implanted state. There are actually two as-implanted states: one is paracrystalline, and the other shows no medium-range order. Both states are unstable, so they can be thermally annealed to a relaxed state. We believe that the fully relaxed state is the continuous random network.; The results of post-anneal implantation also indicate that the degree of medium-range ordering depends on ion mass. More specifically, the degree of medium-range ordering increases with ion mass and energy. This has been confirmed by the observations of depth dependence and the effect of implantation. We speculate that the origin of the paracrystalline state during implantation is associated with “energy spikes.” This model can also qualitatively explain the dependence on ion mass and energy. Our results have an important implication for the processing of silicon by ion implantation, and further computer simulations of this fascinating phenomenon will be very helpful.