About one week before IBM celebrated its 100th year, IBM researchers published in the journal Science [10 June 2011, paywalled, Wafer-Scale Graphene Integrated Circuit] and publicly announced the design of a high speed graphene circuit. Since there are announcements about this or that new application of graphene just about every week, it would be easy to take the IBM announcement in stride and basically ignore it as a run-of-the-mill piece of science or technology. That would be a mistake.
Three words stand out that elevate the importance: graphene, circuit and IBM. Graphene as you probably have heard by now is the non-new pure carbon material with unsuspected properties that were, in part, made practical to the world of research as recently as 2004 (and resulted in a Nobel Prize for the effort). Since then, the pace of research and application development has been nothing short of astonishing. Significantly, one of the leaders in that research has been IBM. [SciTechStory: Graphene transistors] IBM was among the first to produce a working transistor using graphene (2009-2010), which at the time was considered difficult because graphene is not naturally a semiconductor (unlike silicon, for example). However, even the first working graphene transistor IBM built was already twice as fast as a comparable silicon transistor. That meant full speed ahead, in more ways than one.
As one IBM engineer put it, a working transistor is nice but it means nothing until it’s connected to something. That was the next step, which IBM just announced. That’s where the word circuit comes in, specifically in what is called an integrated circuit (IC) meaning that many transistors are connected to perform a task. In this first instance, IBM scientists at the Thomas J. Watson Research Center (New York, USA) constructed what is called a broadband radio-frequency mixer, used in radio applications to process signals at different frequencies. It’s a standard IC component, and while what IBM built what is essentially a ‘proof of concept’ device, it demonstrates that graphene transistors can be integrated (which was not easy).
That’s where IBM manufacturing expertise becomes crucial. Proof of concept and prototypes in the laboratory are one thing, commercial products in mass production are quite another. While IBM is not saying it can mass produce graphene ICs, this announcement makes that implication plausible. IBM solved some of the trickiest technical problems such as protecting the ultra thin (1 atom thick) layer of graphene during the process of etching with electron beam lithography – a standard industry process but applied to graphene instead of silicon.
Ultimately what may be important is that IBM is apparently gaining confidence in graphene as a potential or even probable replacement material for silicon. So far, the FET (Field Effect Transistor) made of graphene can be faster and the integrated circuit it makes can be more flexible and heat resistant than silicon. With the predicted upper limit of speed for silicon approaching within the decade, graphene is a powerful candidate for a replacement. The stakes are huge, nothing short of a new generation of electronic components. If IBM is the first to make graphene transistor devices in mass production; that is almost certainly where the industry will go.