IBM Patented a One Atom-Thick Graphene Transistor That Works 1,000 Times Faster Than Silicon

IBM Patented a One Atom-Thick Graphene Transistor That Works 1,000 Times Faster Than Silicon:

Next stop, the singularity. IBM has just won a patent for a brand new single atom graphene transistor that will allegedly “transmit electrical pulses 1,000 times faster than silicon.” And if that doesn’t give rise to ultra-intelligent, super-efficient autonomous A.I., then man, I don’t know what will.
The tech blogs are heralding the patent as a lifeline for Moore’s law, which for decades now has correctly described the fact that the number of transistors we can fit on a computer chip has doubled every two years, allowing them to get faster, faster.
IBM took to its Tumblr to celebrate the patent, giving itself a pat on the back/PR blast and claiming that this “could give a jolt to the rate that our computers and electronics improve—and uphold Moore’s Law for decades to come.”
Here’s the abstract of patent #8,344,358 in full:

A graphene-based field effect transistor includes source and drain electrodes that are self-aligned to a gate electrode. A stack of a seed layer and a dielectric metal oxide layer is deposited over a patterned graphene layer. A conductive material stack of a first metal portion and a second metal portion is formed above the dielectric metal oxide layer. The first metal portion is laterally etched employing the second metal portion, and exposed portions of the dielectric metal oxide layer are removed to form a gate structure in which the second metal portion overhangs the first metal portion. The seed layer is removed and the overhang is employed to shadow proximal regions around the gate structure during a directional deposition process to form source and drain electrodes that are self-aligned and minimally laterally spaced from edges of the gate electrode.

And one of the renderings filed:


Figure 1 is a "vertical cross-sectional view of an exemplary structure after formation of a graphene layer," which would appear to be the top layer, while figure 1A is a bird-eye view of the same structure. It appears that IBM expects a graphene chip to be a multi-layer device–which makes sense considering the graphene layer is only an atom thick–but it's in the graphene layer that the magic happens. The properties of graphene allows for the creation of what are called graphene field effect transistors, which (this is hopelessly simplified, of course) are powered by self-aligned electron gates that can work at extremely high frequencies—as high as 100 GHz, according to the filing, although there are still technical challenges to get to that point sustainably.
Nonetheless. I am suddenly painfully aware of how crappy and insufficient my 2010-era processing power is, now that I know it could be going 1,000 times faster. I am still aware of page load times, and it is already 2013. What a gyp.
Most of us can hardly even conceive of what a computer that fast might possibly be able to do, especially for humble consumer purposes. Guess we’re going to find out, tho—fingers cross for something closer to legit AR utility on mobile devices or gonzo VR fun and further from autonomous robo-domination. Either way, things are going to get interesting.

Top image