Study of the nucleation mechanism of carbon nanotubes by field emission techniques

Ni was, in the end, the only metal studied, due to the poor quality of images acquired from Co and Fe. Aimed at reproducing the conditioning step of the catalyst often observed in CVD protocols, a first study showed that the crystal adopts a polyhedral morphology at the working temperature (873K) in an hydrogen atmosphere or under Ultra-High-Vacuum conditions, by the extension of dense crystal planes like {111} or {100}. The presence of hydrogen in the chamber does not seem to present any influence on the final crystal morphology at temperatures above 600K.

When exposed to a carbon-containing gas, nickel crystals present two distinct behaviors following the temperature region that is explored. At temperatures below ~623K, exposing Ni to ethylene or acetylene leads to the formation of a stable and poorly structured nickel carbide layer. The superficiality of this carbide is proven by the ease of its physical (by increasing the electrical field) or chemical (exposure to hydrogen or oxygen) evacuation. These three treatments initiate a clean-off phenomenon that evacuates the carbide layer. Reproducing these experiments in the atom-probe confirmed the carbidic nature of the surface as NiCy compounds were collected.

At temperatures above 623K, the carbide layer (formed by exposing Ni to the same gases) becomes unstable. Its formation is related to a transition period that precedes the nucleation of graphenes on the surface. The Ni crystal undergoes a massive morphological transformation when acetylene is introduced in the chamber at 873K. This phenomenon is induced by the presence of carbon on the surface which adsorbs so strongly on step sites that it provokes their creation. Carbon also induces a considerable enhancement of Ni atoms mobility that allows for this transition to occur. Once the new morphology is attained, nucleation of graphenes is observed to start on the extended and carbon-enriched step-containing crystal planes. By reproducing these experiments in the atom-probe, a high surface concentration of carbon dimers and trimers was observed. A kinetic study of their formation was thus achieved and showed that they were formed on the surface by the recombination of Cad. Their potential role as building-blocks of the CNT growth process (which had previously been proposed following theoretical considerations) is thus suggested on the basis of experimental results for the first time.

Two critical surface concentrations are highlighted in the present work. The first one is needed for the formation of carbon dimers and trimers and the second one has to be attained, during the morphological transformation, before the onset of graphene nucleation, probably providing a sufficient growth rate of the graphitic nuclei and allowing them to attain their critical size before their decomposition.

Finally, the observation of rotational circular patterns, most probably related to carbon nanotubes, suggests that CNT growth (and not only graphene nucleation) occurred episodically in our conditions, confirming the validity of our model.