News & Views | Published:

Nanotechnology

Electronics and the single atom

Naturevolume 417pages701702 (2002) | Download Citation

Subjects

The invention of semiconductor transistors in the 1940s revolutionized electronic circuitry. In the new world of 'nanoelectronics', a transistor whose active component is a single atom has now been demonstrated.

Nanotechnologists are seeking to build nanometre-scale electronic devices in which the functional unit is a single molecule or atom1. Carbon nanotubes have proved particularly useful as molecular 'wires'2 in this quest, thanks to their long lengths (by nanotechnology standards at least) of several micrometres. So could it be possible to wire up a short molecule, or even a single atom, to create a nanoscale transistor? The experiments reported by Park et al.3 and Liang et al.4 (on page 722 and page 725 of this issue) give a positive answer.

Trapping a molecule between two metal electrodes to make such a transistor is a tough technological challenge1,3,4,5,6,7. First, the molecule needs suitable terminations that reliably bind it chemically to the two electrodes, bridging the gap between them. Conventional lithographic techniques by which the electrode structure might be assembled have a resolution, at best, of around 10 nm — yet the electrodes would need to be just 1 nm apart. The two groups3,4 have used the unconventional combination of electron-beam lithography and electromigration to reach this 1-nm scale. Still, the entrapment of a single molecule in the electrode gap is an occasional, lucky event. Moreover, at present there is no viable imaging technique for directly confirming successful trapping. Yet the presence of one, and only one, molecule can be indirectly established from its conduction properties.

The molecules used by Park et al. and Liang et al. in their ultra-small transistors are organic complexes that contain one3 or two4 atoms of a transition metal. The metal atoms, cobalt3 and di-vanadium4, form the active region of the device, whereas the organic molecule serves as mechanical support and provides the connection to the metal electrodes. Both experiments demonstrate transistor operation based on a tunable flow of electrons through the metal atom. (Although this simple picture is largely correct, the properties of the metal atom are strongly affected by the presence of the organic molecule.)

The current through an electronic transistor can be turned on and off by changing the voltage on a gate electrode. In the 'on' state, current is carried by a large number of electrons, with typically a billion passing per second — the large number is necessary to obtain a measurable current. In a commercial silicon transistor, electrons move independently of each other by diffusive motion from the 'source' to the 'drain' terminal. Although the motion of an individual electron cannot be predicted, the average motion of a large number of electrons can, resulting in well-defined transistor operation.

How do electrons flow through a single-molecule transistor? The answer is by a simple fundamental process. The repulsive Coulomb interaction between electrons means that there is an energy cost in adding an extra electron to any object of small dimensions (this same energy cost is in part responsible for the ionization and affinity energies in an atom). In a molecular transistor, this energy cost can be tuned to zero by applying a voltage to the gate electrode. At a particular gate voltage, the electrostatic potential is such that an extra electron can hop from the source onto the molecule. However, Coulomb repulsion forbids a second, extra, electron hopping on at the same time — the first electron must leave the molecule, moving into the drain, to make way for the next electron. This one-by-one electron motion, governed by the quantum of electron charge, is known as single-electron tunnelling.

But there is another quantum property important for electron transport through small objects — the electron spin. Electrons each carry one half-unit of spin in one of two configurations, defined as 'up' or 'down'. As there is no preference for the direction of the spin, the lowest-energy (ground) state of the system is a combination of up and down spins. To reach this ground state, an electron spin must flip between up and down, and this can be arranged by replacing an up-spin electron on the molecule by a down-spin electron from one of the electrodes. This spin-flipping, driven by the desire of the system to be in its ground state, enforces an exchange of electrons between molecule and electrodes through a process called the Kondo effect8.

In the experiments by Park et al.3 and Liang et al.4, the molecule is tuned between two states that differ by one charge and one spin unit. In one state, the conduction mechanism is single-electron tunnelling. In the next state, with an odd number of electrons on the molecule, electron transport is mediated by the Kondo effect. Remarkably, when the Kondo effect dominates, the conductance reaches values close to the quantum limit (which is 2e2/h, where e is the electron charge and h is Planck's constant) for a perfectly conducting one-dimensional wire. This result is quite remarkable, considering the intrinsic difficulty of establishing good electrical connection between a molecule and a metal electrode.

Do these realizations of a single-atom transistor mean that molecular electronics is just around the corner? That goal may be a little closer, but there is still a long road ahead before atomic or molecular transistors can be assembled into viable, dense, fast logic-circuits. Right now, these single-molecule or single-atom transistors are no competition for silicon transistors. But they will serve both scientifically, for studying electron motion through nanoscale objects, and technologically, for developing chemical techniques with which to fabricate electronic devices on single molecules.

References

  1. 1

    Joachim, C., Gimzewski, J. K. & Aviram, A. Nature 408, 541–548 (2000).

  2. 2

    Dekker, C. Physics Today 52, 22–28 (1999).

  3. 3

    Park, J. et al. Nature 417, 722–725 (2002).

  4. 4

    Liang, W., Shores, M. P., Bockrath, M., Long, J. R. & Park, H. Nature 417, 725–729 (2002).

  5. 5

    Park, H. et al. Nature 407, 57–60 (2000).

  6. 6

    Reichert, J. et al. Phys. Rev. Lett. 88, 176804 (2002).

  7. 7

    Zhitenev, N. B., Meng, H. & Bao, Z. Phys. Rev. Lett. 88, 226801 (2002).

  8. 8

    Kouwenhoven, L. & Glazman, L. Physics World 14, 33–37 (2001).

Download references

Author information

Affiliations

  1. the ERATO Mesoscopic Correlations Project, Department of NanoScience, Delft University of Technology, PO Box 5046, Delft, 2600 GA, The Netherlands

    • Silvano De Franceschi
    •  & Leo Kouwenhoven

Authors

  1. Search for Silvano De Franceschi in:

  2. Search for Leo Kouwenhoven in:

About this article

Publication history

Issue Date

DOI

https://doi.org/10.1038/417701a

Further reading

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing