Jun 16 2011

Graphene Breakthrough: Things Are Only Going To Get Faster From Here


It seems like every time I turn around there’s a new “breakthrough” in graphene news.

Don’t know what graphene is? Well, I’ve written about it before over at H+ magazine and R.U. asked me to write a follow up following the latest bit of news that IBM has built a complete circuit — a broadband frequency mixer — from graphene. For those of you not familiar with electronics, a broadband frequency mixer is one of the most basic of integrated circuits, one that converts a signal from one frequency to another, i.e. audio range frequencies into radio, or even the different frequencies of audio in a equalizer. As a demonstration of the utility of graphene in electronics, only making a functional logic circuit would be more impressive.

When I wrote about Graphene for H+ and its potential for computers, people working in the field hadn’t yet come upon a couple of hurdles that needed to be overcome. The first was that graphene has a low bandgap, which basically means that you can’t fully turn off a graphene transistor the way you could a silicon one. In some applications, like radio electronics, this would be fantastic. But for computers, it was a huge problem, because all logic circuits are basically on/off switches. Some people thought that would be a killer problem, but it turned out to be resolvable. This is because graphene is a very remarkable material.

Graphene is composed of a single sheet of atoms, being essentially a two dimensional plane, so it has some very different properties than most 3 dimensional materials. For example, the shape of a piece of graphene effects its electrical properties. If it’s bent, or wrinkled, those properties change, and even the shape of a “trace” (those flat ribbon wires that you see on any circuit board) effects it. By making a squared off U-bend shaped transistor, the bandgap problem suddenly disappeared, enabling a bandgap nearly a thousand times higher than all previous attempts to make transistors. They still don’t know why changing the shape makes such a dramatic change, but this increase basically eliminated the worries about graphene digital circuits.

But that wasn’t the only worry.  Researchers discovered that their original notion —  that they could use graphene the same way copper was used in integrated circuits — wasn’t going to be feasible. The same effect that solved the bandgap issue caused a problem when layering graphene on silicon. The imperfections of the silicone surface created “pools” of low energy states, so rather than traveling through the graphene like it had when the graphene layer was isolated from a surface, electrons just kind of settled into the pools and stayed there.

In many ways, electricity acts like water, so it makes a handy visualization tool. Silicon is a crystal matrix, and small imperfections in this crystal structure create areas of localized positive and negative charges. Positive charges act like “hills,” while negative charges act like “valleys.” When single layer graphene was layered on silicone, it conformed to the “hills” and “valleys” of the silicon, rather than maintaining its “flatness.” And because shape has such a powerful effect on graphene, these hills and valleys introduced “resistance” into graphene that wasn’t there in its isolated state.

Now, this wasn’t a “killer” problem either because it would have eventually been possible to create all-graphene circuits, but it did appear to be a damper on the immediate applications of graphene in current electronics, because it eliminated one of graphene’s biggest advantages — its ability to conduct electricity, nearly like a superconductor, at room temperatures. The fact that it got far worse in the presence of a magnetic field, causing electrons to basically come to a standstill, seemed to be a barrier that would be difficult to overcome.

While they were looking for solutions to these issues, researchers discovered that molybdenum had many characteristics in common with graphene and, unlike graphene, was also a semiconductor. It also had a bandgap much higher than had been possible with graphene, prior to their discovery of something called the squared u trick. But research into molybdenum’s potentials was far behind the research on graphene, meaning it would likely take years for researcher’s to learn enough about its properties for I to be very useful.

Fortunately, right around the time the discoveries about molybdenum hit the news, the discovery about the U shaped bend transistor came out, followed by news that layering graphene on boron nitride virtually eliminated the problem with the “pooling” electrons by eliminating the “wrinkles” Soon thereafter, researchers discovered that graphene transistors are self-cooling

In earlier articles for H+ on this topic, I discussed how the lower resistance of graphene was likely to allow the creation of processors that ran thousands of times faster than current ones, but with the same amount of electricity and heat generated.  Now it appears that graphene chips will actively cool themselves. In other words, unlike silicon, the chips will not need cooling.  This promises to eliminate one of the biggest headaches in electronics.

Heat control is a significant factor in both size and power consumption in nearly all electronics devices. A self-cooling integrated chip could be miniscule, since they will no longer require the plastic “cases” that move heat out of the actual chip and transfer it to a cooling system. Even circuit board design will be effected.  We will no longer have to place heat producing components far apart so they don’t overheat each other.

With regard to the news of IBM’s frequency mixer, one little detail immediately caught my eye: “The researchers demonstrated the circuit at up to 10GHz, and showed this level of performance was stable at up to 127°C.” That’s 260.6F, people!… far past the point where your computer processor overheats and goes into thermal breakdown. Taken together with the news about graphene’s self-cooling properties, it shows that graphene is likely to be functional in nearly any environment we can put it in — great news for a future of rugged portable electronics. But to get to the real news, you have to read a little further. It’s this line: “The method will work with existing optical lithography, IBM said, and can be applied to graphene films created by chemical vapour deposition. This means existing fabs would not need massive revamping to start using graphene in earnest.”

That’s the real point you need to understand. My little history lesson above is all about illustrating the importance of these words, and the real point that IBM is making.

I ended my earlier H+ article with this line: “In other words, graphene could begin making its way into computers as early as 2012 to 2015, and almost certainly by 2020.”

IBM couldn’t wait. This is only the beginning, and merely a proof of concept, but the gauntlet has been thrown. Graphene has come out of the lab and into the real world.

Welcome to the GHz Wars Version 2.0.  Things are only going to get faster from here.