New Graphene Research Promises Reliable Chip-Level Production

An anonymous reader writes "A research team from the University of Texas and a German nanotechnology company have published a paper which describes a major milestone for the future of graphene-based computing – the reliable production of wafer-scale graphene measuring between 100 and 300mm, suitable at last for integration with 'traditional' materials in computing. The research team was able to manufacture 25,000 graphene field-effect transistors from lab-produced graphene film on a polycrystalline copper base. Team research leader Deji Akinwande said: 'Our process is based on the scalable concept of growing graphene on copper-coated silicon substrates...Once we had developed a suitable method for growing high-quality graphene with negligible numbers of defects in small sample sizes, it was relatively straightforward for us to scale up.'"(Original, paywalled paper is at ACS Nano.)

Comments

Interesting.

[Anonymous Coward]

I wonder if this will help me 3D print a drone? I was originally going to use Raspberry Pi, but might use a graphene based product instead. I would like to start a delivery service which accepts bitcoin as payment for Apple products.

Re:Interesting.

[NoKaOi]

I wonder if this will help me 3D print a drone? I was originally going to use Raspberry Pi, but might use a graphene based product instead. I would like to start a delivery service which accepts bitcoin as payment for Apple products.

I think you could do it, but you'll need to use nanotechnology.

Re:Interesting.

[CajunArson]

Can't decide... legitimate troll or bot designed to get +1 votes by vomiting buzzwords and hoping for the best...

Re:Interesting.

[penguinoid]

It's legitimate satire. His "business plan" would be possible to implement before graphene even comes close to fulfilling its "promise". Not saying graphene doesn't have the potential, it's just that sort of thing takes way longer than journalists are allowed by their editors to admit.

Article written by clueless PR bots

[CajunArson]

The non-paywalled article includes some hilarious zingers like "the material also has extraordinary semiconductive properties which could revolutionise the issue of cooling in data centres."

If by "extraodinary" you mean: No bandgap unless you are really doctoring the graphene with other materials, then sure since "ordinary" semiconductors have bandgaps.

Not sure how transistors that can't be turned off will help in cooling data centers, but who knows what revolutions lurk in future press releases!

Re:Article written by clueless PR bots

[ShanghaiBill]

No bandgap unless you are really doctoring the graphene with other materials

Bilayer graphene has a bandgap, and the bandgap can be tuned by applying a voltage, applying strain, etc. Graphene is not a drop in replacement for silicon, and a lot of work needs to be done before it shows up in commercial products. But it has some very interesting properties, and may eventually replace most silicon electronics.

Not sure how transistors that can't be turned off will help in cooling data centers

Well, I don't know about "cooling data centers", but it should result in cooler semiconductors. Graphene conducts heat better than silicon, and if it is built on a copper substrate, with just a few carbon atoms separating the heat generation from the copper, it should be much easier to dissipate heat. That means it can either run cooler, or run much faster.

Re:Article written by clueless PR bots

[Anonymous Coward]

Graphene semi-conductors can run both cooler and much, much faster from what I've seen. There are also different substrates to grow it on, which one will be best is a question.

But there's little question that Graphene could well replace silicon for processors soon enough. Between much higher electron mobility, and thus much higher clockspeeds, better quantum tunneling behavior than silicon allowing Moore's law to continue beyond 7nm feature size, and the seemingly endless interest in the material from corporations and universities I'd bet graphene based processors (and other parts of a computer) will be here within a decade or sooner

Re:Article written by clueless PR bots

[SacredNaCl]

I'm reasonably familiar with the technologies currently used in the chip making process (and always walk away impressed with what they can do), but I don't see how graphene is that superior of a material? What advantages does it really have? A smaller lattice? Can other known materials used in the chip making process truly use a smaller lattice or are we running up against the end here? Or is it just that its lattice has 6 atoms instead of 4 that is has the advantage for trying to put in more transistors on a chip (with a far less straightforward design process)?

I realize the test process is going to be pretty low density until they figure out how to remove it, but I'm unfamiliar with any other advantages.

 

Re:Article written by clueless PR bots

[smaddox]

Graphene exhibits very high thermal conductivity in plane, but that doesn't help much with getting heat out of a chip, into a heat-sink and then into the air.

Graphene as a switch makes very little since (the lack of a bandgap leads to high leakage/poor modulation depth), but there is a chance it will replace or augment metallization in silicon chips. The push to smaller metallization lines has caused several reliability issues which are still not completely solved. Graphene has shown some promise in this area.

Re:Article written by clueless PR bots

[smaddox]

It is difficult to imagine how graphene could replace silicon CMOS when the best lab MOSFETs can't even compete with mass produced Si MOSFETs. High mobility doesn't matter when the off-state leakage current is a couple orders of magnitude higher than in Si. There are interesting theoretical proposals for bi-layer graphene switches which could compete (e.g. BiSFET), but no experimental demonstration that I'm aware of.

Re:Abstract from paper published ACS Nano:

[Required Snark]

This is significant progress, but "better" does not mean good enough. Consider the following:

"150 and 300 mm wafers reveals >95% monolayer uniformity". If I understand current process yield amounts, a single layer that is only 95% good is not sufficient for large scale manufacturing. You need lots of layers, and the yield goes down with the product of the percentages for each layer.

"26000 graphene field-effect transistors were realized". On 100mm or 300mm? Compared to current density for VLSI it's many orders of magnitude off.

"About 18% of devices show mobility of >3000 cm2/(V s), more than 3 times higher than prior results ". Wonderful, but prior art was only 6%. What are the values for the other 82%? Are they useful at all?

"polycrystalline graphene". Are their any currently deployed polycrystalline graphene transistor devices? I thought that to make graphene really shine it needs to be more uniform then polycrystalline, hence more a single crystal structure.

So if you think that university press releases are bullcrap, then wanna-be company press releases are significantly lower on the food chain. Cockroach crap? Flea crap?

wafer-scale

[Anonymous Coward]

I am not sold on wafer-scale until they make it web-scale