Silicon memristors promise cheaper, faster, denser memory
First silicon oxide resistive ram operates at room temperature, cheap as chips.

University College London researchers have detailed what it claims to be the world's first silicon oxide-based resistive RAM (ReRAM) that can run at room temperature.
ReRAM is one of the technologies researchers are developing that could replace traditional NAND flash for non-volatile storage systems, offering speeds far closer to that of volatile dynamic RAM (DRAM) modules.
Previous ReRAM prototypes, including a headline-grabbing example from 2008 which was based on titanium dioxide and Eplida's recently-announced module which is due for mass production in 2013, have required esoteric operating environments - including vacuums and extreme temperatures - and complex production processes which have precluded them from commercialisation.
According to the team's work, published in the Journal of Applied Physics, the solution may have been found with the first ReRAM module to be produced from common silicon oxide and to operate at ambient temperatures - heralding a potential revolution in high-speed storage systems.
"Our ReRAM memory chips need just a thousandth of the energy and are around a hundred times faster than standard flash memory chips," boasted Tony Kenyon of his team's findings. "The fact that the device can operate in ambient conditions and has a continuously variable resistance opens up a huge range of potential applications."
The prototype system is capable of switching resistance significantly more efficiently than any other ReRAM prototype around, thanks to a novel silicon oxide structure whereby the arrangement of silicon atoms changes to form filaments of silicon within the solid silicon oxide. These filaments are significantly less resistive than the filament-free silicon oxide, providing the one to their absence's zero required for digital storage.
The ReRAM prototype could also be used to produce memristor hardware which mimics the neurons in the brain, thanks to its ability to record a continuously variable resistance based on the last voltage that was applied.
Amusingly, the material was discovered by accident. During work on producing silicon-based light-emitting diodes (LEDs), researchers noted that the prototype devices appeared unstable. Adnan Mehonic, a PhD student, was asked to investigate - and found that, far from being unstable, the material the team had created flipped between conductive and non-conductive states extremely predictably.
"My work revealed that a material we had been looking at for some time could in fact be made into a memristor," Mehonic explained at the material's unveiling. "The potential for this material is huge. During proof of concept development we have shown we can program the chips using the cycle between two or more states of conductivity. We're very excited that our devices may be an important step towards new silicon memory chips."
The team is keen to point out the material's potential for high-speed non-volatile memory systems, memristor applications and even for use as a central processor - but, as is often the case with academia, is somewhat silent on a potential release date for a commercialised version of the technology.
Meeting the future of education with confidence
How the switch to digital learning has created an opportunity to meet the needs of every student, always

The Total Economic Impact™ of IBM Cloud Pak® for Watson AIOps with Instana
Cost savings and business benefits
