News & Views | Published:

Device physics

Silver nanoswitch

Ionic conductors have many applications — in sensors, fuel cells and batteries. Are nanoelectronic devices based on ionic conductors now about to replace silicon?

Most electronic appliances are based on digital electronics, which in essence just require a lot of switches working together in an organized fashion. Much research has been aimed at finding a reliable switching mechanism that can beat conventional silicon technology to permit ever smaller and more powerful electronics. The ideal switch should be scalable down to atomic size; it should have low power consumption, and require just two leads for both read and write memory operations. On page 47 of this issue1, Terabe and co-workers describe an invention that comes close to this ideal. They exploit the fascinating properties of silver sulphide, a material in which electrical conductivity is carried by both electrons and silver ions. The resulting devices can be used for logic as well as for fast memory operations, and they function at room temperature.

In most solids, atoms sit at fixed positions in a regular crystal lattice. In the solid ionic conductors used by Terabe et al., however, some ions have many possible equivalent positions in the lattice and can wander through the material. Figure 1 illustrates this for the conductor of interest here, Ag2S. When the material is connected by two silver leads to a battery, Ag+ ions are formed at the interface between silver sulphide and the positive silver electrode, while Ag+ is reduced at the other electrode. This process leads to the transport of silver, removing it from the positive lead and depositing the same quantity at the negative lead. Ag2S is one of a rare kind of solid ionic conductors that have two unusual features: it operates at room temperature and it conducts electrons as well as ions. Both features are of central importance to the device created by Terabe and co-workers.

Figure 1: Silver sulphide — a mixed electronic and ionic conductor.
figure1

Two silver contacts, at the top and bottom, are applied to Ag2S and connected to a battery. The current is partly carried by electrons, partly by positive silver ions (circled ‘plus’ signs) diffusing through the sulphide in the opposite direction. The ions are replenished at the positive electrode by oxidation of the electrode material, while silver is reduced and deposited at the other end.

A few years ago, the authors reported that nanoscale silver mounds formed on top of a Ag2S crystal when a scanning tunnelling microscope (STM) was used2,3,4. In that experiment, a silver bottom electrode is held at a positive electrical potential with respect to the platinum STM tip. Electrons tunnelling from the tip to the surface of Ag2S are partly used up in reducing Ag+ ions to metallic silver. Keeping the tip at a fixed height above the surface results in the formation of a silver metallic bridge between tip and sample. The process can be reversed by reversing the electrical potential, which dissolves the silver bridge back into the sulphide. This is the principle of the switch: contact can be made or broken by applying a voltage of the appropriate sign. The reason this work went largely unnoticed is that many switching mechanisms between STM tips and substrates have been discovered in recent years, but they are of little practical value because each device requires its own STM. For practical applications, such tunnel junctions between two electrodes need to be controlled in a simpler and more reproducible way.

Terabe et al.1 have come up with a clever solution: they exploited the properties of the ionic conductors themselves to create and control the required tunnel gap. A layer of Ag2S on top of a silver wire is in contact with a thick platinum wire through a silver layer one nanometre thick (Fig. 2a). The platinum and silver leads are then connected to a voltage source to run an electronic current from top to bottom. This current is accompanied by the transport of silver downwards through the silver sulphide, and after a few seconds the top silver layer vanishes, resulting in a loss of contact with the platinum lead. The device is now in the ‘off’ state and ready for operation. When the polarity of the applied voltage is reversed (Fig. 2b), a local silver bridge is promptly formed which again closes the gap between the platinum and Ag2S, turning the switch ‘on’. The process can be reversed and repeated rapidly because only a few atoms are involved.

Figure 2: A rewritable memory bit based on the properties of the silver sulphide (Ag2S) mixed ionic conductor, as described by Terabe and colleagues1.
figure2

a, A one-nanometre-thick silver layer deposited on top of the Ag2S layer disappears into the sulphide layer when a current flows from the platinum (Pt) lead to the silver lead. This results in loss of contact between the two electrodes and initializes the device. b, A bridge of silver atoms is locally formed by applying a voltage of opposite sign, re-establishing contact between the silver sulphide and platinum. The conductance through the device can be as small as one quantum unit of conductance, suggesting that the silver bridge can touch the platinum lead with just one atom.

Terabe et al. observe, moreover, that the conductance of the device can be as small as one quantum unit if a short voltage pulse of the correct amplitude and duration is applied. In this case, it seems that the silver bridge has grown upwards until just one atom touches the platinum lead (for a review, see ref. 5). To switch between on and off states requires voltages higher than 100 mV. The state of the memory bit — that is, whether on or off — can be read non-destructively at voltages lower than that, taking advantage of the electron-conducting property of the silver sulphide.

By combining two silver sulphide switches with resistors and capacitors, Terabe et al. carry out the basic logic operations AND, OR and NOT. In principle, this is all that is needed to perform more complex logic operations. However, the efficiency of these logic gates will be significantly reduced when their inputs and outputs are connected to other logic gates in a large digital circuit. In order to avoid this problem, there should be some way of amplifying the gate signals; otherwise, their logic applications will unfortunately be limited. The multiple steps in quantum conductance that the authors observe with increasing voltage are also remarkable, but probably not of sufficiently practical use in view of their limited reproducibility.

Yet the main result is of great beauty and simplicity, and is scalable to nanometre-sized addressable bits. The authors have done well to protect their work with several patent applications.

References

  1. 1

    Terabe, K., Hasegawa, T., Nakayama, T. & Aono, M. Nature 433, 47–50 (2005).

  2. 2

    Terabe, K., Hasegawa, T., Nakayama, T. & Aono, M. RIKEN Rev. 37, 7–8 (2000).

  3. 3

    Terabe, K., Hasegawa, T., Nakayama, T. & Aono, M. Appl. Phys. Lett. 80, 4009–4011 (2002).

  4. 4

    Terabe, K., Hasegawa, T., Nakayama, T. & Aono, M. J. Appl. Phys. 91, 10110–10114 (2002).

  5. 5

    Agraït, N., Levy Yeyati, A. & van Ruitenbeek, J. M. Phys. Rep. 377, 81–279 (2003).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

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.