A silver lining

Luminous silver nanoclusters could hold the key to the next generation of computers, reports Jon Evans.

After holding true for almost 40 years, Moore’s law could soon be consigned to oblivion. Postulated in 1965 by Gordon Moore, co-founder of Intel, the law predicts that the number of transistors that can be fitted onto an integrated computer chip will double in each technology generation (roughly every 18 months).

In line with this law, the number of transistors on a chip has increased from 30 in 1965 to a few hundred million today. Now, however, Moore’s law is beginning to run up against the laws of physics.

Moore’s law has held true because the semiconductor industry has continuously managed to shrink transistor dimensions, but this cannot continue indefinitely. Already, computer scientists are encountering problems with fabricating electronic components and controlling the supply of electrons, due to the ever decreasing scales.

To overcome these problems, researchers are looking at novel ways of designing and making computer chips. These include building nanoscale electronic devices from the bottom up, and developing optoelectronic circuits, which use photons rather than electrons to transmit information. However, the optimum solution could be to combine these two approaches, and this is exactly what two chemists from the Georgia Institute of Technology, Atlanta, US, have now done.

The chemists, Robert Dickson and Tae-Hee Lee, have produced arrays of silver nanoclusters that generate light when an electric current is applied. And because the nanoclusters only generate light when a current of the right voltage is applied, each nanocluster is able to act as a primitive computer logic gate, performing AND, OR, NOT and XOR functions. ¹

Linking these nanoclusters together should allow them to perform more complex computing functions. What’s more, because the nanoclusters form spontaneously, work at room temperature and don’t need individual electrical connections, they overcome many of the problems that plague current nanoscale electronic devices.

Dickson first came across these silver nanoclusters ² when investigating fluorescence in films of silver oxide (Ag ₂O). He and his team of researchers discovered that silver oxide films, although initially non-fluorescent, would start to fluoresce in a variety of colours when illuminated with wavelengths of light shorter than 520nm. Dickson attributed the fluorescence to small silver nanoclusters that form on the silver oxide film as a result of the irradiation. The variety of colours produced could be explained by the fact that the irradiation was constantly changing the size and charge of the nanoclusters.

Dickson then teamed up with Lee to explore the potential of these silver nanoclusters. ³ They discovered that the clusters, which are made up of between two and eight atoms, would also form when an electric current is passed through a silver oxide film for several seconds. The current induces ’electromigration’, which creates a nanoscale break junction in the most resistive part of the film; the nanoclusters form along the break. When an ac current is subsequently applied to the break junction, the clusters generate light, producing a variety of different colours that span the visible spectrum.

This electroluminescence is a result of the electric current injecting holes (extracting electrons) into the nanoclusters and then injecting electrons, which recombine with the holes and produce photons. The researchers discovered that the electron-hole recombination time that produced the most powerful electroluminescence was of the order of a few nanoseconds; any longer and the holes would begin to decay thermally. They also found that the different clusters, being a variety of sizes and possessing different energy levels, would only electroluminesce at certain defined voltages.

Dickson and Lee worked out that they could control the electron-hole recombination process in the silver nanoclusters by applying two short pulses of electric current. The first - a positive pulse - inserts the holes, and the second - a negative pulse - inserts the electrons, thereby causing the nanocluster to electroluminesce. However, the first and second pulses must be at exactly the right voltages to generate any light. By varying the first and second pulses, a nanocluster is therefore able to act as a logic gate.

For instance, one nanocluster might photoluminesce only if it receives a +2.2V first pulse followed by a -1.05V second pulse. If the second pulse is a combination of two pulses that can each be either 0V or -1.05V, then the nanocluster will act as an XOR gate, ie it will only photoluminesce if the second pulses add up to -1.05V, not if the pulses add up to 0V or -2.1V. By combining the pulses in this way, single nanoclusters can act as AND, OR, NOT and XOR logic gates, with light as the output.

Furthermore, because each nanocluster only responds to very specific voltages, individual nanoclusters can be controlled without needing isolated electrical connections, which are hard to generate at the nanoscale. Another advantage of the nanoclusters is that they show clean on-off behaviour, because exactly the right voltage is required for luminescence. It is becoming increasingly difficult for modern transistors to obtain high ’on’ currents and low ’off’ currents, because of their small size.

Dickson and Lee claim that their silver nanoclusters could allow more complex calculations to be performed with vastly simpler circuits than are possible with standard transistors. ’Many people are trying to shrink electronics down to the nanometre scale and take advantage of the interesting properties that arise when you make things very small’, says Dickson, ’but often they are using standard architectures to create logic circuits. We are using a novel architecture to do the same thing’. Moore’s law could continue to hold true for a while yet.

Source: Chemistry in Britain

Acknowledgements

Jon Evans