February 07, 2011

Towards Mastery of Carbon Nanotube Growth

Carbon nanotubes could enable plastics to be manufactured that are ten times stronger than the strongest materials available today. Mastering carbon nanotube growth would enable high volume and lower cost production of longer and stronger carbon nanotubes so that the vision of using carbon nanotubes for more applications in society can be realized. Currently only about one thousand tons per year of carbon nanotubes are produced.

Here is a 109 page thesis on Carbon Nanotube growth

Materials used above 1 billion tons/year

carbon 8.6 billion tons/year
iron 1.4 billion tons/year
cement 3.1 billion tons/year (calcium carbonate)

Materials used above 100 million tons/year

Phosphorous 153,000,000 Tons/Year
calcium 112,000,000 tons/year
oxygen 100,000,000 tons/year

Materials used above 10 million tons/year

Sulfur 54,000,000 Tons/Year
hydrogen 50,000,000 tons/year
nitrogen 44,000,000 tons/year
potassium 36,000,000 tons/year
aluminum 30,000,000 tons/year
copper 15,000,000 tons/year
zinc 12,500,000 tons/year

Glass fiber production (used for reinforcement of plastics and cement) is at ten million tons per year.

Carbon fiber production is at about 70,000 tons per year.

Carbon nanotubes production at tens of millions of tons per year at lower cost than glass fiber would transform current glass fiber using applications.

Carbon nanotube production of 100,000 tons per year at lower cost and higher quality than carbon fiber would transform the current markets using carbon fiber.

Conclusions and Outlook of the CNT thesis

The work presented in the thesis has focused on a number of parameters that are of importance in the search for methods of controlled growth of CNTs. The size of the systems under study has spanned from single particles to infinite CNT forests, i.e., from the single metal clusters present prior to CNT growth to the large CNT ensembles present after growth

Although the temperature during CNT growth is significantly lower than the bulk melting temperature of the metal used as catalyst these metal clusters may be liquid. This is attributed to the known size dependence of the melting temperature but also on effects induced by external parameters. For clusters mounted on a substrate both the adhesion strength between the substrate and cluster and the surface structure of the substrate were seen to affect the melting temperature. For CNT growth this means that the same catalyst metal may be either liquid or solid depending on which substrate it is mounted. This can provide a partial explanation for the observed effect of the substrate on CNT growth.

In the studies of the attachment of CNTs to metal clusters it was found that CNTs
are attached above the cluster surfaces and that clusters adapt to the shape of the CNT and not vice versa. The effect of this is that the cluster may be geometrically deformed while the CNT will retain both shape and chirality at the end. This trend was observed irrespectively of whether the cluster was a pure metal or a metal carbide. The metal carbide systems were however less stable than the corresponding systems with pure metals although the the adhesion strength between CNTs and carbide clusters was comparable to the adhesion strength between CNTs and pure metal. This means that if the adhesion is strong enough for pure metal clusters to remain attached to the CNT it is likely that also carbide clusters remain attached to the CNTs. The lower stability of the carbide systems may however imply that the CNT drain the clusters from excess carbon atoms.

Upon variation of the the size of the metal clusters attached to the CNTs it was found that a minimum cluster size is needed to prevent the CNT ends from closing. This is a possible explanation for the observed trend that CNTs grow from cluster particles with diameters that are similar to or slightly larger than the diameters of the CNTs. From these results it could be concluded that both pure metals and metal carbides may be docked onto open CNT ends in order to achieve continued growth of CNTs as well as for the creation junctions between CNTs and electrode materials. The exact behaviour of these systems may however depend on the size of the clusters, the carbon content and the temperature as the mere possibility of docking is not a guarantee for the successful achievement of neither continued growth nor operational contacts.

The learnings from these studies were transferred to studies of large ensembles of CNTcluster systems. This was done in order to understand if the Ostwald ripening of the catalysts may provide any information of relevance for the controlled CNT growth. The systems under study mimicked the experimental distributions of CNTs with different diameters and chiralities attached to clusters of different size. Both the adhesion strength and the CNT diameters were seen to affect the properties of the Ostwald ripening of the clusters. It was predicted that clusters attached to CNTs of large diameters and strong CNT-cluster adhesion are most likely to survive Ostwald ripening. This means that these CNTs should have a longer growth time than other CNTs although introduction of radicals in the growth chamber may impose radical changes to the trends.

Finally, a simple phenomenological model for simulations of the Ostwald ripening in cluster ensembles was constructed. The model reproduced the stationary LSWdistribution describing Ostwald ripening as well as the qualitative trends of the DFT calculations for nano clusters attached to CNTs. The model did also allow for predictions of the long time behaviour of the CNT-cluster distributions after Ostwald ripening.

Although these studies have provided information that take the CNT community a few
steps further towards the controlled CNT growth much work is likely to remain before this becomes reality. The next study to be performed within the project will address the stability of C atoms on the surface of metal clusters. Hopefully this study can provide information about both the formation of the CNT cap and the mechanisms during CNT growth. This study is most likely to be conducted with DFT and TBMC calculations in a combination that will allow for high accuracy, flexibility, finite temperature and hopefully simulated CNT growth. The structure in Fig. 7.1, showing a CNT that has continued growth from an existing seed CNT, indicate that the latter may become reality in the ”near” future. Preliminary results indicate that C atoms are more stable when forming strings on the cluster surface compared to attaching to the cluster surface as monomers. Hence, the early formation of strings and polygon structures on the cluster surface appears to be more likely than random positioning of C atoms on the surface. Another indication is that C atoms in the middle of the strings may detach from the surface which may be important for the rejection of the cluster from the CNT end. It may also limit the maximum allowed length of the chains before soot is formed instead of CNTs.

If you liked this article, please give it a quick review on ycombinator or StumbleUpon. Thanks
blog comments powered by Disqus