Graphene: Nanoelectronics goes flat out
Marcus Freitag is at the IBM TJ Watson Research Center, Yorktown Heights, New York 10598, USA.
The unique electronic band structure of graphene has led to a number of exotic effects that have fascinated fundamental researchers and may also lead to improvements in the performance of electronic devices.
IntroductionWhen drawing a line with a pencil, small fragments of graphite are left behind on the paper, but it wasn’t until four years ago that researchers discovered that single sheets of carbon atoms called graphene could exist among the debris. They went on to demonstrate that this two-dimensional material had electronic properties quite unlike those observed in most other systems. Since then the number of papers on the subject has mushroomed and shows no sign of slowing down as researchers explore new ways of producing graphene and exploiting its remarkable electronic properties in devices.
Graphene is a semiconductor but it has the unusual feature that its band gap is exactly zero. Moreover, the velocity of the charge carriers in graphene does not decrease at the top of the valence band and the bottom of the conduction band, as is usual for most materials, but instead stays constant throughout the bands, including at the Dirac point where the conduction and valence bands meet (see Box 1). A gate voltage can, however, modulate the density of states in graphene and switch between the low-conductivity state at the Dirac point and the high-conductivity states elsewhere.
Such a switching action is at the heart of the field-effect transistor — the central element in modern computer chips — which is why there is so much interest in graphene, and its close relative, the carbon nanotube, for applications in nanoelectronics. However, there is a crucial difference between the two: semiconducting nanotubes can have band gaps of about 1 eV, which effectively blocks the current in the off state, whereas the low-conductivity state in graphene carries a finite current, even at temperatures close to absolute zero. The minimum conductivity in graphene is also compromised by defects, impurities and the substrate on which it has been placed.
Now, Eva Andrei and co-workers at Rutgers University1, reporting on page 491, and independently, Philip Kim and colleagues at Columbia University2, 3 have explored the intrinsic electronic properties of graphene close to the Dirac point by suspending graphene devices to remove the influence of the substrate (Fig. 1). The Columbia group has also removed contaminants in a current-induced thermal annealing process. All substrates can trap electric charges, producing image charges in the graphene. One consequence of this is the formation of electron–hole ‘puddles’ where, for gate voltages close to the conduction minimum, parts of the graphene sheet act as electron conductors, while other parts of the same sheet act as hole conductors.
Changes in the gate voltage can redistribute the charges in the puddles, but they can not reduce the conductivity below a certain value, which depends on the amount of disorder. (Again, graphene is unusual in this regard because high levels of disorder would cause the conductivity to drop to zero in most materials). In suspended graphene, on the other hand, the minimum conductivity at the Dirac point approaches a universal (geometry independent) value of 4e2/h at low temperatures1, where e is the electron charge and h is Planck’s constant, as predicted by several theories. Moreover, the transition between high and low-conductivity states as a function of gate voltage becomes very sharp, and the charge density at the Dirac point drops to below 1010 electrons per square centimetre (which is just one electron in an area of 100100 nm), making it conceivable to study, in the future, the last few electrons that are free to move.
Both groups have also measured electron mobilities in excess of 105 cm2 V-1 s-1 at carrier densities above 1011 cm-2, which is an order of magnitude improvement over substrate-supported devices. At room temperature, mobilities are closer to 104 cm2 V-1 s-1, limited by acoustic phonon scattering, but this is still much better than current silicon devices. The reported mobilities imply that electrons can travel from one contact to the other with only a few scattering events in the channel or at the channel boundary. This could one day enable ballistic graphene electronics to be used for high-frequency applications.
Source (and continue reading this paper): http://www.nature.com/nnano/journal/v3/n8/full/nnano.2008.219.html