Samsung Claims Breakthrough In Graphene Chip Design

jfruh (300774) writes "Graphene, a carbon-based crystalline lattice that is extremely strong, lightweight, and an excellent conductor of electricity and heat, is coveted as a potential base for semiconductor chip design, and Samsung, working with the Sungkyungkwan University School of Advanced Materials Science and Engineering, has claimed a big jump towards that goal. With IBM also making progress in this realm, the days of silicon could actually be numbered."

Producing them is one thing

[Viol8]

Producing them cheaply enough to rival chips made of processed sand is another matter entirely. Anyone remember gallium arsenide chips that were going to eat silicon for lunch back in the 80s? Yeah , well.... still niche.

Don't Worry

[The Cat]

Some asshole on Slashdot knows why it will never work.

Was this a PR statement or major breakthrough?

[GodfatherofSoul]

That seems like an odd announcement to make...if it's just one more step in the research process and this doesn't make graphene a viable replacement *yet*...

Re:Producing them is one thing

[AmbiLobe]

The expansion of carbon does not match the expansion of insulators when the temperature changes. Silicon matches the thermal size changes of silicon dioxide. If Samsung has matched the coefficients of expansion, it is big news. But that was not announced.

Re:Producing them is one thing

[K. S. Kyosuke]

Silicon has always been cheaper, not just recently. Any fancy substrate or switch technology will be unusable as a replacement for silicon if it can't at least approach the density of silicon circuits. We can manufacture extremely complex circuits using the very simple procedures of silicon crystal growing, doping, photolithography, and epitaxy. The moderate cost of modern ICs in the middle of the synchronous price-per-component curve (as per Moore's law) is the result of the combination of these simple procedures (which are ridiculously simplistic when it comes to simple components) with doing them at extreme geometrical complexities (which is *the* thing that makes ICs expensive). Even if you could demonstrate a single graphene switch today, you would still be where Jean Hoerni's planar silicon transistor was in 1960. Now find out how this supposedly very tough material can be either deposited and grown into proper shapes or substractively machined at nanoscale, and all that at high speeds needed for mass IC manufacture.

Incompetent article writer?

[K. S. Kyosuke]

I think that an article whose author claims that "Germanium ... doesn't occur naturally" and that "400Ghz ... should make for some strong signals" ought to be taken with a very large lab-grown monocrystal of salt.

Seriously, this is great news!

[sumakor]

Graphene needed this technological development. It was a pre-requisite for electronics applications, which are currently based on large single crystal silicon wafers. For comparison, this is something that's yet to be achieved with carbon nanotubes, which still have no electronics applications despite being 13 years older than graphene and having excellent properties. People have the same attitude towards graphene: yeah it's great, but it may never be integrated into any mass-produced products and it may just die out and fade away. So if Samsung can grow monocrystalline graphene many inches across, it moves graphene from some pie-in-the-sky research material like nanotubes to something we could actually commercialize. It knocks out one of the big legs from the "Graphene will never replace silicon," argument. Although not all the questions about graphene have been answered, this advance makes those questions and their answers matter a lot more to many more people than they did last week.

Re:Producing them is one thing

[wagnerrp]

The problem is that transistors are thermoelectric devices. You switch them on and off by heating them up to change their conductivity. Silicon chips can withstand temperatures well beyond the point at which the plastic packages they are mounted to break down, but that temperature is also well beyond their switching point, making them useless as a computational device.

If you could produce a semiconductor that was useful at 3000F, then that would be its normal operating temperature, and you would need to feed it a high enough core voltage to allow it to heat itself up to that temperature to switch.