Next-generation memory devices that use phase change materials at the nanoscale have taken a step closer to realisation thanks to research performed at the University of Pennsylvania in Philadelphia, USA.
Phase change materials offer the potential to create faster and longer lasting memory devices, due to their ability to quickly switch from a low resistance crystalline phase to a high resistance amorphous phase, allowing them to read and write data swiftly. But creating such electronic instruments at the nanoscale using traditional lithography processes has been difficult. Lithography damages the materials’ surfaces and interferes with data storage capabilities.
The research team at Pennsylvania has used a non-lithographic method to grow phase change nanowires. Powder forms of germanium, antimony and tellurium are heated until they vaporise. The vapours are run over a piece of silicon studded with gold nanoparticles, and as they cool, the nanoparticles act as catalysts, forming wires that grow to between 30-50nm in diameter and 10nm in length.
Memory devices are created by assembling strands of these nanowires on a silicon substrate and defining the electrical contacts. During testing, researchers found that the wires are around 1,000 times faster than flash memory and should be able to hold data for 100,000 years.
Ritesh Agarwal, Assistant Professor of Materials Science and Engineering at the University of Pennsylvania, says the nanowires also consume less power (0.7MW) than current memory devices.
‘Most nanoscale memories are based on poorly characterised phenomena, slow switching speeds and do not show size-dependant properties,’ says Agarwal. ‘In our nanowires, the mechanism is based on well characterised structural transformations, and we have demonstrated all the attributes of an ideal memory device.’
‘This approach of creating nanowires by catalysed growth is well known,’ comments Professor Jeremy Allam, Deputy Director of the Advanced Technology Institute at the University of Surrey, Guildford, UK. ‘What they’ve done is show that you can use a bottom-up technique to produce structures which, in principle, could be used in a phase change memory. The memory properties are rather nice, but there are still technical barriers.
‘The challenge [will be] to grow them in a controllable way, [including] both the size and shape of the wire and its position. And that challenge is generic across all bottom-up nanotechnology techniques, which prevent them, at the moment, from being used in real devices.’
The Pennsylvania team is investigating methods of assembling the nanowire devices on a large scale, as well as studying the physics of nanoscale transition phenomena in nanowires. They are also looking into the limits of size, speed and data-retention on such devices.