Analysis of size-distribution function of metallic nanoclusters in hydrogenated amorphous carbon matrix

A method is proposed for mathematical processing of TEM images of nanocomposites with a large amount of non-uniform noise. The approach is 2D spline filtering of the image using the values of average background intensity. The method has been applied to TEM images of amorphous carbon doped with copper. A size distribution function of copper clusters has been obtained which is in a good accordance with earlier data. A dependence of cluster-to-cluster distance on cluster size has been found. This dependence validates the hypothesis of fluctuation growth of the clusters at lower metal contents and their coalescence at higher contents.