Focus on Carbon Electronics: Graphene – Nanotubes – Diamond

Authors

  • John Robertson


Carbon has three allotropes, each of which has been a star material at the time. Graphene, consisting of single layers of sp2 carbon, is presently the most exciting form of carbon, ever since it was made by the mechanical exfoliation of graphite by A. K. Geim et al. 1. Its unique band structure of mass-less Fermions due to the Dirac cone structure, and the possibility of huge carrier mobilities with inhibited carrier scattering has led to intense fundamental research activity. The gap-less band structure requires the two carbon atoms in the planar unit cell to be symmetry equivalent, unlike in the Bernal graphite structure.

Graphene is currently limited by large area production. The most mature method for producing large-scale samples is presently the sublimation method from the hexagonal faces of SiC. The Si atoms sublime, leaving one or more surface layers of carbon atoms. The group of W. A. de Heer 2 have pioneered this method for the C-terminated (0001) face. They have described how rotations of the graphene plane above the carbon buffer layer directly bonded to the underlying silicon atoms, allow the graphene to retain the required equivalence of its carbon sites 3. Other production methods include chemical vapour deposition, particularly on copper, where deposition self limits to a single layer 4, and solution based methods via graphene oxide or via chemical-induced exfoliation 5.

These achievements have led to unique conducting properties such as the Berry phase and Landau levels observable at room temperature 6, and extremely large, frequency-independent optical absorption of 2.3% per atomic layer.

The high mobility in conjunction with a vanishing band gap means that graphene will be less useful for logic transistors which must be turned off, but will be more relevant to high frequency analogue RF transistors where fT's of order 50 GHz are already possible 7. The high mobilities arise because important carrier scattering mechanisms are symmetry forbidden. It is of interest to understand how to minimize the remaining scattering processes such as scattering by charges in any underlying insulator. Here, Raman scattering can be useful in identifying the dominant mode, as in the Rapid Research Letter by C. Casiraghi 8. The fabrication of transistors requires lithographic definitions, and novel chemical patterning methods are interesting, such as that of E. Bekyarova et al. 9.

Before graphene, carbon nanotubes were the heavily researched form of carbon. Whereas graphene has one of the simplest lattices, carbon nanotube research was maintained because of the complexity of unravelling the different behaviours of tubes of different chiral index. Nanotubes also have very high mobilities and make high performance transistors. Much of the characterisation of the different types of nanotubes has used luminescence or Raman spectroscopy 10, 11. It was interesting that for many years the electronic spectra of nanotubes were described in terms of the independent electron density of states with van Hove singularities. It should have been clear that their one-dimensional structure leads to low screening, and thus a dominance of excitons in their optical properties 12. This approach has been continued in the Rapid Research Letters of E. Malic et al. 13 and J. Lefebvre et al. 14.

Nanotubes, fullerenes, and graphene are forms of sp2 carbon. In 1988, the most exciting form of carbon was diamond, the sp3 allotrope, because for the first time it could be synthesised as thin films by chemical vapour deposition 15. It was Science's ‘molecule of the year’. Diamond has the most extreme properties of any three dimensional crystal, in terms of atomic density, hardness, Young's modulus, and thermal conductivity. Because of its three dimensional crystal structure, diamond devices might be made using traditional semiconductor processing techniques of growth, doping and etching. However, diamond's extreme bond energy and large diffusion energies meant that it is extremely difficult to process dopants by diffusion, to anneal out defects, and it is mainly a unipolar semiconductor. These problems have delayed applications of this material and limited them to important but niche areas. The article by K. Haenen et al. 16 highlights how to overcome some of these problems and achieve a bipolar device. Interestingly, considerable recent work in diamond now involves nanocrystalline diamond which is smooth, and can be used as a substrate for new directions in chemical and biochemical functionalization for sensors.

Thumbnail image of

John Robertson is a Professor at the University of Cambridge. He has a wide interest in all forms of electronic materials, and has worked on diamond, carbon nanotubes, graphene and amorphous carbon. He is a member of the Editorial Board of physica status solidi – Rapid Research Letters (pss-RRL).