Researchers from Rice University have recently published the results of their work on graphite-based mass data storage and reprogrammable gate arrays in a major step towards making graphene-based electronics a reality.
Using graphite to build the hard drives of tomorrowGraphite has long been known to have unique electrical properties and has therefore been put forward by many as a possible substitute for silicon for use in integrated circuitry.
The Rice University research team, led by Prof. James Tour, deposited several 10-nanometers-thick stripes of amorphous graphite onto silicon and later verified that this process facilitated the creation of potentially very dense and stable non-volatile memory.
For reasons that are yet to be completely understood, graphite behaves in a very peculiar way whenever a current is run through it. When a current is run though a 10-atom-thick layer of graphite, a small gap — only two nanometers in size — is immediately formed, effectively breaking the circuit in two electrically insulated parts.
What's even more surprising, when a current is run again through the same circuit, the break is instantly repaired. This process can apparently be repeated indefinitely, providing a simple and yet highly efficient way of representing a bit.
The graphite-based approach holds a number of notable advantages over current technology such as Flash memory, including a vastly increased memory density; a low operational voltage of only three volts; a very high on/off charge ratio, facilitating the bit "read" process; the need for only two terminals instead of the usual three, which greatly simplifies the circuitry needed; and finally, its high resistance to temperature changes and high radiation levels, which make it suitable for deployment in the space and military industry.
Slashing the costs of circuit design
Non-volatile memory is not the only possible application — perhaps not even the most promising one — for the team's work.
When embedding vertical stripes of graphite in "vias," the holes that connect the different layers of an integrated circuit, an antifuse is formed. Antifuses are the basic elements of a common kind of FPGA, a type of "blank" circuit that can be programmed via software. Typical antifuse FPGAs can only be programmed once but, harnessing the same break-and-repair mechanism seen in mass memory, using graphite can make them reprogrammable at will.
A reprogrammable antifuse FPGA could prove extremely helpful for chip designers. Right now, to build their circuits, fabricators need to rely on elaborate photolithographic masks that are becoming increasingly expensive as electronics keeps getting smaller and denser. The cost of designing and manufacturing such a mask has increased almost tenfold during the last decade, now reaching millions of dollars.
Reprogrammable antifuse FPGAs would however mean that designers will be able to experiment new designs and make modifications on-the-fly without significant costs, accelerating innovation in the sector and eventually bringing more efficient chip designs on the market.
Via Rice University.