Modulating the height of carbon nanotube forests by controlling the molybdenum thin film reservoir thickness.

Many catalyst materials have been tried to synthesize ultra-long carbon nanotubes (CNTs) by extending catalyst lifetime and delaying growth termination. We propose a time-controlled, variable composition iron-molybdenum catalyst system, where the diffusion of molybdenum (as a thin layer reservoir) is mediated by the alumina underlayer, to reach and to slowly alloy with the Fe catalyst on the surface during the thermal process. This technique enhanced both the catalytic activity and the catalytic lifetime to grow CNT carpets with heights up 5 mm, compared to a maximum of approximately 1.5 mm for a regular sample (without Mo reservoir). Moreover, the CNT height increased with the thickness of the Mo thin layer reservoir for thicknesses from 10 nm to 30 nm. We discuss this new growth mechanism using high resolution transmission microscopy (HRTEM) images of cross-section lamellas and Rutherford Back Scattering (RBS) analysis to show the increasing alloying of Mo with Fe. Overall, the proposed technique of mediated diffusion of Mo to the surface with subsequent progressive alloying with the Fe catalyst, besides enhancing CNT height, could allow the one-step synthesis of CNT carpets with regions of different heights based on patterning these regions with different thicknesses of the Mo reservoir during sample preparation.

What is below the support layer affects carbon nanotube growth: an iron catalyst reservoir yields taller nanotube carpets.

Here we demonstrate an approach to enhance the growth of vertically aligned carbon nanotubes (CNTs) by including a catalyst reservoir underneath the thin-film alumina catalyst underlayer. This reservoir led to enhanced CNT growth due to the migration of catalytic material from below the underlayer up to the surface through alumina pinholes during processing. This led to the formation of large Fe particles, which in turn influenced the morphology evolution of the catalytic iron surface layer through Ostwald ripening. With inclusion of this catalyst reservoir, we observed CNT growth up to 100% taller than that observed without the catalyst reservoir consistently across a wide range of annealing and growth durations. Imaging studies of catalyst layers both for different annealing times and for different alumina support layer thicknesses demonstrate that the surface exposure of metal from the reservoir leads to an active population of smaller catalyst particles upon annealing as opposed to a bimodal catalyst size distribution that appears without inclusion of a reservoir. Overall, the mechanism for growth enhancement we present here demonstrates a new route to engineering efficient catalyst structures to overcome the limitations of CNT growth processes.