Quantitative electron energy loss spectroscopy of Si nanoclusters embedded in SiO x

Abstract

We have used a methodology, based on electron energy loss spectroscopy combined with energy filtered images, which allows us to quantify the clustered silicon concentration in annealed sub-stoichiometric silicon oxide layers (SiO x ). This information was coupled to the chemical silicon and oxygen concentrations determined from Rutherford backscattering analyses. The silicon agglomeration kinetics was investigated as a function of the gas flows used during the plasma enhanced chemical vapor deposition process. Although the clustered Si concentration and the average cluster radius do not follow a well defined dependence on Si, O, or N concentration, separately, the ratio between the former parameter and the silicon excess concentration decreases monotonically as a function of the nitrogen concentration approaching a saturation value equal to 0.2 for nitrogen concentrations above 10 at.%.