Smaller is almost always better in the world of computer technology, but there are technical limits. The components are getting so tiny that existing manufacturing techniques can't easily take miniaturisation any further.
Consumers had got used to a doubling of computer power every six months with each new generation of microchip, but this would not continue, according to Prof Donald Fitzmaurice who heads the Nanochemistry Group at University College Dublin. "That will just stop happening," he said. "After 2012 it is generally accepted that greater miniaturisation will not come from existing technologies."
The answer, he believes, is to "grow" rather than build the next generation of computer hardware. "One of the things we are trying to do is to grow chips in a beaker."
Growing smaller components had great potential, he said. With self-assembly you get "structures which build themselves". His team has developed a process that can be used to chemically grow gold wires no more than 12 nanometres (a nanometre is a billionth of a metre) across - about one eight thousandth of the width of a human hair.
The work is published this morning in the journal, Advanced Materials, and involved Mr Stephen Fullam, Dr David Cottell, Dr Hakan Rensmo and Prof Fitzmaurice, all of the Nanochemistry Group. The smallest microchip transistors now measure about 200 nanometres across, and are produced by lasers. These incredibly tiny junctions are lumbering giants, however, compared to the fine nanowires being produced at UCD.
The group was looking for ways to produce nanowires by growing them in solution. "Recently we have been playing around with a new approach using carbon nanotubes," Prof Fitzmaurice explained. Nanotubes are an exotic form of carbon related to fullerenes - remarkable football-shaped lattices formed by 60 interlocking carbon atoms which were only recently discovered. Fullerenes, and the hollow, tubeshaped lattice the nanotube, are produced by high-voltage discharge. Depending on the conditions you either get fullerenes or nanotubes. Prof Werner Blau of Trinity College produces these structures and he supplied the UCD team with the nanotubes used in its experiments.
The group wanted to use gold to coat the nanotubes and so produce an ultra-fine wire that could conduct electricity. It first produced gold particles about six nanometres across and then coated them with a chemical, tetractylammonium bromide. "We made particles which are coated with a molecule that likes to stick to the nanotubes."
This is done in solution, with the gold particles covering the tubes as soon as they are mixed in. A method was then needed to fuse the particles into a continuous, stable surface and this was achieved by heating.
"It is like a kitchen foil roll covered in marshmallows," Prof Fitzmaurice explained. "You put it into the oven and they all fuse together, only you are doing that on a billionth of that scale."
The heating makes use of a phenomenon called "skin melting". If a surface is curved you can get melting of a substance at lower temperatures, and in this case the gold particles fuse at between 250 to 300 degrees Celsius, compared to about 1,000 degrees for ordinary gold.
"We get a continuous gold sheath around the nanotube," he said, producing a wire about 12 nanometres across and about 10,000 nano metres long. The work was important he said, not just because the wire is very small but also because it could be made using self-assembly in solution.
These fine wires are now being tested at other research labs to see how well they carry an electric current. Some have also been sent to an IBM research lab where they are being studied for their "quantum effects". The wires are so small they can be influenced by other nearby atoms.