Growing nanowires on an insulating support is a key step forward in nanodevice fabrication since it ensures that conducting nanostructures are isolated from the substrate. This means that no current leaks from the nanowires into the substrate.

Geoff Thornton of University College London and colleagues at the London Centre for Nanotechnology, and the University of Manchester grew palladium nanowires in ultrahigh vacuum by metal vapour deposition of a palladium wire from a directional source. The metal wire is bombarded with energetic electrons that heat the wire and cause it to melt, generating a metal vapour. The substrate is held at 900K during metal deposition and no further annealing is required.

"Our metallic nanowires self-assemble in a single crystallographic direction that is determined by the anisotropic surface structure of the substrate, rutile TiO2(110)1x1," lead author David Humphrey told nanotechweb.org. This directionality may allow the researchers to use templated thin films of TiO2(110) to control the growth direction of the nanowires in devices.

Unprecedented aspect ratios
This high-temperature growth mode has never been observed on metal oxide surfaces before and metallic nanowires with aspect ratios of around 1:200 (nanowires just 5 nm wide and 1000 nm long) are unprecedented, added Humphrey.

The UCL team analysed the wires at the Nanoscience Beamline I06 at the Diamond Light Source (DLS) in Oxfordshire. The scientists employed Low Energy Electron Microscopy for initial characterization and X-ray Photoemission Electron Microscopy (XPEEM) to monitor the chemical state of individual nanoparticles. In this way, they determined that the palladium nanoparticles were indeed metallic and growing on a dielectric substrate, rutile TiO2(110).

DIAMOND facility
DLS is a new third-generation, high-brightness synchrotron light source. It has the selectivity and surface sensitivity required for the UCL study and enables short acquisition times. For example the secondary electron XPEEM image of the Pd nanowires took only 30 seconds.

Achieving the result was no easy task though. Thornton and colleagues had to find the right temperature range to grow the nanowires, since lower temperatures lead to small elongated structures. "The growth conditions also had to be well defined so that they were reproducible and transferable," said Humphrey. "This is important as recreating preparation conditions from one experimental chamber (that at UCL) to another (DLS) can get complicated."

The fact that nanowire growth is highly directional could come in very useful for fabricating real-world devices. For example, the long unidirectional wires could be used as interconnects for nanoscale electronic circuits, such as high-density computing components or in micro- or nano-scale electromechanical systems, providing power or signal channels. TiO2 is also suitable as a gate dielectric in field effect transistors (FETs) and the palladium nanowires could be used as interconnects in such devices.

The researchers say they would now like to unravel the detailed mechanism for nanowire growth, since it is as yet unknown. "If we could identify the nucleation process and specific nucleation sites could be fabricated, this would allow us to control the growth of the nanowires," explained Humphrey.

The team is also involved in growing thin films of single-crystal TiO2 on metal substrates and modifying the surface chemistry and structure of single crystals of TiO2 with the aim of patterning/templating. "Again, this will be useful for controlled nucleation and growth of the palladium nanowires."

The work was published in Nano Letters and will shortly be presented at the 17th International Surface Science Conference at the University of Reading (30 March – 2 April 2009).