The role of ion flux in Si(100) nanostructuring by normal-incidence Ar+ ion sputtering has been studied. The measured relationships of the Si lateral dot size versus ion flux, the surface roughness versus ion flux, and the surface roughness versus sample temperature with ion fluxes of 20 and 380 µA cm−2 all indicate that the value of the ion flux is decisive for the validity of the Bradley–Harper (BH) model in the nanostructuring of semiconductor single crystals. In this work, for Ar+ ion sputtering of Si(100) with ion energy of 1.5 keV, it is found that only beyond ∼220 µA cm−2 is the BH model well applicable, while below that the Ehrlich–Schwoebel (ES) one tends to be involved. Our results suggest that the ES barrier effect is negligible under relatively high flux conditions, while it is substantial in the case of relatively low flux; for the BH model, the situation is just the reverse. Hence, caution should be exercised as regards the value of the ion flux when one tries to tune the semiconductor nanodot size following the BH model.

Photoluminescence (PL) from SiOx (0

The effects of CeF3 doping on the photoluminescence (PL) of Si nanocrystals embedded in a SiO2 matrix, assigned as nc-Si:SiO2, have been studied. The nc-Si:SiO2 sample was prepared by means of reactive evaporation followed by thermal annealing, and CeF3 doping was conducted via thermal diffusion. It is found that the doping of CeF3 is able to enhance the PL intensity of nc-Si by a factor as large as 3.7. With the increasing doping concentration, the degree of PL intensity enhancement increases until a maximum is reached, and then it drops down. This trend also holds for the PL intensity enhancement versus the diffusion annealing temperature. A continuous redshift of the PL peak was observed with the increasing doping concentration and diffusion annealing temperature. Beyond a certain concentration or temperature, the PL peak starts to shift back towards shorter wavelengths. The doping effects on the PL intensity were explained by a model of electron transfer from Ce ions to nc-Si, while the trends of PL peak shift were accounted for with the help of a previous model referring to the size distribution of nc-Si crystallites versus nc-Si:SiO2 film thickness (Fang et al 2004 Nanotechnology 15 494); the peak shift was also related to the chemical reaction between Ce and Si.

No abstract

We report an ion flux dependence study of the Si dot pattern formed on Si(100) by Ar+ ion sputtering with the ion energy being 1.5 keV, ion dose 5 × 1017 ions cm−2, and ion flux ranging from 280 to 1100 µA cm−2. Experimental results show that the lateral dot diameter d and the ion flux f basically follow the relationship of , and the surface roughness w decreases with increasing f in an exponential decay manner. Simulations based on a widely accepted continuum model, namely the noisy Kuramoto–Sivashinsky equation, reproduced the trend for d versus f but failed to explain that for w versus f. A redeposition consideration was then suggested. It is found that with this correction not only are the d–f and w–f relationships well explained, but the simulated surface morphology bears closer resemblance to the experimental one as well. The effect of redeposition becomes important for f> ∼130 µA cm−2 as derived in this work.