IBM has made an announcement in the journal Science – arguably the United States’ finest scientific publication – stating that they have found a material for their computer chips that will leave the silicon-based devices far, far behind. Their new transistors are made from graphene, which is a material gaining huge amounts of attention due to its unique structure and extremely promising characteristics.
Graphene is related to words you already know, such as graphite. Graphene is a single sheet of carbon atoms, all lying in the same plane, and connected by sp2 bonds to form a honey-comb type arrangement – almost like molecular chicken-wire. I’ve attached a picture of a graphene structure to this article. In it’s bulk form – with many layers of these sheets all stacked on top of each other – graphene is known as graphite, and isn’t very interesting. Graphite can be used as a lubricant, because the sheets can easily slide past each other, becoming quite slippery. However, its bulk form is not of any interest for use in electronics.
However, in 2005 it was discovered that if a single sheet of graphite – known as a graphene sheet – could be isolated by itself, and separated from the rest of the pile, it became enormously interesting for its electronic behavior. It became a “semi-metal”, which means it is still technically a non-metal (it’s made up of nothing but carbon atoms, and carbon isn’t something that we normally characterize as being a metal), but it conducts electricity just as easily as a metal. The electron mobility – the ease with which electrons can move through the material – was very promising, and researchers immediately began to test its other properties. Amazingly enough, while graphite is of no use as a structural material – the individual sheets slide past each other with ease, providing little resistance – a single graphene sheet is enormously strong: more than 200 times stronger than steel.
This is an amazing material, and it has now been put to use by IBM to make transistors which switch on and off 100 billion times each second, which is ten times faster than the fastest silicon-based transistors. These new transistors were created using methods which are compatible with those used in everyday semiconductor plants. This means that the probability of this new technology being adopted is very high; manufacturers won’t have to install brand new fabrication equipment or become accustomed to entirely new ways of operating.
Previously, researchers made graphene transistors using laborious mechanical techniques, such as flaking off sheets of graphene from graphite. Mechanically separated graphene is one of the most expensive materials on the planet, and was not commercially viable. Graphene however is very attractive in that its base cost – just the carbons – is very cheap; it’s separating the material and isolating the individual sheets that is expensive. While very fast transistors can be made from exotic semiconducting materials such as indium, they are never going to be a commercial success due to the high cost of the starting materials.
The breakthrough for IBM came about by using a sicon-carbide wafer as the base for their transistors, the so-called “dielectric”. By heating the wafer slowly, the silicon began to evaporate, which left beyind a thin layer of carbon in the form of graphene. This has huge advantages for processing as instead of trying to pluck a single sheet out of a slippery mountain of graphite, and then trying to manipulate that sheet (which is only one atom thick) and lay it down on a surface, all the researchers have to do is take a block of silicon carbide (a very tough, manageable material) and heat it under the right conditions. The silicon goes away, and the carbide that is left over is transformed into the graphene structure, which is already “in place” and right where it needs to be.
Graphene transistors have enormous potential because the electrons move so much faster through graphene, as compared to silicon. Even without doing any optimization of the design, already these new carbon-based transistors are 2.5 times better than their silicon counterparts. We should see these new materials start to take over for silicon in high-speed computer processors relatively soon, and they will provide a very noticeable and affordable speed boost to our electronics.