We have investigated experimentally and theoretically the role that ripening processes play in the evolution of supported transition metal nanoclusters on solid surfaces. We discuss some avenues of analysis that can be more informative of the dominant mechanisms than mere temporal measures of island size. Our model system is formed by the growth of Pd nanoclusters at room temperature on the cross-linked (1 × 2) reconstructed TiO2(110) surface. The reconstructed surface contains a rectangular array of 'defects' that nucleate and pin the clusters. We have followed the surface evolution by variable temperature scanning tunnelling microscopy as the temperature was raised stepwise to 973 K. We find that the ripening is dominated by an Ostwald-type mechanism with particles remaining immobile during growth (decay). Monte Carlo simulations of the ripening of arrays of three-dimensional islands have been undertaken for comparison with the experiments. These reproduce the spatial properties of the experimental arrays and the scaling nature of the height distribution. In addition the classic mean field theory for the size distribution is modified to include island size-separation correlations and is found to recover the size distribution found in both the simulation and the experiment. We conclude that detailed atomistic understanding of ripening is not always necessary to understand important features of nanostructure evolution.
- surface diffusion