DISORDERING OF PLATINUM CLUSTERS

Platinum clusters less than 4 nm in diameter appear to ''melt'' when exposed to the electron beam in a scanning transmission electron microscope. The diffraction patterns of these disordered clusters show dynamic features which suggest that some micro-crystalline structure may still remain. The calculated steady-state temperature increase due to the interaction of the electron beam with the cluster and substrate is not sufficient to cause the cluster to melt even with melting-point depression considered. The clusters remain amorphous for an average of 20 minutes on a carbon substrate, and for 24 hours when supported on silica. The cause of ''melting'' (or disordering) and the recrystallization mechanism are, at present, unknown.