Relation between metal electronic structure and morphology of metal compounds inside carbon nanotubes

International audienceSEVERAL attempts have been made to fill carbon nanotubes(1) with metals or metallic compounds to obtain nanocomposite materials with potentially interesting properties. Capillary action, predicted(2) to be a filling mechanism, has been used(3,4) to encapsulate lead and bismuth in open tubes. Compounds of yttrium(5), manganese(6) and gadolinium(7) have also been encapsulated by formation of the nanotubes in an are discharge with the metals present in situ. Very recently, Tsang et al.(8) showed that oxides of nickel, cobalt, iron and uranium can be encapsulated by opening the tubes and depositing the filling material using wet chemical techniques. Here we report a search for general principles relating to the nature and structure of the filling material, using the are-discharge method to fill tubes with fifteen metals and/or their compounds: Ti, Cr, Fe, Co, Ni, Cu, Zn, Mo, Pd, Sn, Ta, W, Gd, Dy and Yb. We find that the propensity for forming continuous 'nanowires' throughout the length of the tubes seems to be strongly correlated with the existence of an incomplete electronic shell in the most stable ionic state of the metal. We also find that the interplay between growth of the nanotube and growth of the filling results, in one case, in the formation of an unusual helical filling morphology