Phaedon Avouris, an IBM fellow and manager of the Nanometer Scale Science and Technology program at IBM's T.J. Watson Research Center, is in the vanguard of experimental and theoretical research into the electrical properties and transport mechanisms of carbon nanotubes and other nanostructures. A comprehensive model of nanotube behavior used by Avouris and colleagues at IBM might serve as a blueprint for the design and fabrication of carbon-nanotube-based electronic devices and circuits. Avouris talked to EE Times' R. Colin Johnson about the surprising operating regions available to designers of carbon nanotube transistors for specific applications.
EE Times: IBM Research has been pioneering carbon nanotube transistors since 1998, when it first fabricated one using a semiconducting nanotube as the channel of an otherwise silicon transistor. Recently, however, you reported on a comprehensive characterization IBM did, which indicates what engineers need to learn in order to design nanotube-based transistors for specific applications. What was the gist of this report?
Phaedon Avouris: [IBM researchers] Vasili Perebeinos, Jerry Tersoff and I have explored in some detail the transport of nanotube field-effect transistors in the finite-carrier-density regime.
EET: Is this the first time that such detailed modeling has been performed for carbon nanotubes?
Avouris: To my knowledge, no one has explored theoretically the finite-density regime. Nothing in silicon devices is done until their performance is first modeled. Unfortunately, we don't yet understand nanotubes as well as we understand silicon. Phonon scattering determines the transport regime in which a carbon nanotube device will operate, and it is very different from that in silicon. It is a field that is very active and evolving, and our work is just our latest contribution to its progress.
EET: How does the finite-density regime compare with the ballistic regime?
Avouris: Nanotubes can be very short or very long, and dif- ferent fields can be applied. If you take a clean, short nanotube, typically 100 to 200 nanometers long, and apply a low bias, the transport--that is, the way the carriers move--is ballistic; there is no scattering from phonons [atomic vibrations within the crystalline lattice] or impurities. We and others have already demonstrated many devices with ballistic transport.
Now we have explored the finite-density regime, where the carrier density is controlled by the gate electrode. For certain electronic devices, you want the nanotubes as short as possible so they are fast. For other types of electronics and for optoelectronics, where you use nanotubes to emit light, you want them longer--on the order of a micron or even longer. What we have most recently explored is the finite-density regime of longer-nanotube transistors, whose transport is called diffusive [as opposed to ballistic], which is the same regime in which silicon transistors operate--so [the comparison is] with silicon.
EET: How was this comparison made?
Avouris: All theoretical modeling up until now has been done assuming an individual electron flowing inside the transistor. In a real transistor, of course, you have switching--that is, you have a third electrode, the gate, that enables you to control the electron density in the transistor's channel [here, a nanotube] by the voltage applied to the gate.