Microchip components are traditionally made by carving patterns into layers of a silicon substrate. In this technique, light is shone through a stencil onto a silicon wafer that has been coated with a light-sensitive polymer known as a resist. A chemical then removes the polymer from certain regions until the desired structure is achieved. However, the resolution of this photolithography technique is limited by the diffraction limit of light, which means that it cannot make features smaller than the current limit of 100 nm.

Jillian Buriak and colleagues at the University of Alberta have now put forward an alternative technique to photolithography that exploits the self-assembly properties of block co-polymers. These polymers consist of long chain-like molecules made from different segments containing two or more monomers. The co-polymers automatically self-assemble into ordered patterns with dimensions of less than 20 nm. These patterns can then be used in the same way as conventional photoresist patterns.

Buriak's team used the block co-polymer polystyrene-poly(2-vinylpyridine), or PS-P2VP, to pattern arrays of lines about 15 nm across that are around 36 nm apart. The material separates out to produce rods of P2VP embedded in a PS matrix when heated at 230 °C for 24 hours. The rods self-assemble to form ordered structures that align along a pre-patterned trench, which is then divided into smaller features (figure 1).

The process does not stop here because the arrays produced can then "build themselves" – a long sought after goal in such fabrication processes. Once the PS-P2VP has self-aligned, Buriak and co-workers use a chemical approach to selectively metallize just the P2VP domains.

They do this by adding acid to the array, which ensures that the P2VP rods are left with a net positive charge. This then attracts negatively charged metal-containing anions when the array is immersed in a metal salt solution. The metal complexes formed are finally reduced to form nanowires and the polymers are removed using an oxygen plasma treatment (figure 2). The process produces arrays of wires that are around 10 nm wide.

The team has already made arrays of gold, platinum and palladium wires and say the technique could be used for other materials, such as magnetic alloys, materials used in the semiconductor industry and organic materials that are important for sensing and other nanoelectromechanical devices.

The team still has lots to keep it busy though. "We will now extend our approach to produce nanowires of other technologically important materials such as copper, iron, cobalt and nickel," team member Steven Chai told nanotechweb.org. "We will also improve the conductivity of these nanowires so that they are of sufficient quality for microelectronics applications."

The work was published in Nature Nanotechnology.