Intel abandons silicon in 7nm process

[CommandMan7]

Intel is saying that 10nm will be the last process to use silicon and is currently investigating alternative materials.

To hit 7nm, Intel says new materials will be required—as in, it looks like 10nm will finally be the end of the road for silicon. The most likely replacement for silicon is a III-V semiconductor such as indium gallium arsenide (InGaAs), though Intel hasn't provided any specific details yet.

While Intel didn't provide any specifics, we strongly suspect that we're looking at the arrival of transistors based on III-V semiconductors. III-V semiconductors have higher electron mobility than silicon, which means that they can be fashioned into smaller and faster (as in higher switching speed) transistors.

This could have deep-reaching implications. Silicon has been used in circuitry and processors ever since Intel released what could be considered the first commercial general purpose programmable microprocessor, the Intel 4004. I'm curious what this could mean for Intel's competition, and what kinds of benefits this could bring. You would assume this would certainly bring a reasonable increase in clock speeds, so I checked a second source;

indium gallium arsenide has long been a promising candidate, however, because its ability to conduct electrons is superior to silicon's at the nanometer scale – about five times better, in fact.

but there's a catch - 

Should they succeed, however, there's still one hitch to overcome before chips based on the new material go mainstream; namely, that indium gallium arsenide – which is composed of the elements indium, gallium, and arsenic – is currently as much as 10 times as expensive as the equivalent amount of silicon. Those chips may be small, all right. But they'll cost you.

The second source is a bit dated, (2012), but hopefully since then the price has gone down. This could revolutionize processing the same way the Intel 4004 did, way back in 1971. Silicon has been the building block of computers for so long, it will be interesting if other chip makers can adapt.
 
 
Sources: 
http://arstechnica.com/gadgets/2015/02/intel-forges-ahead-to-10nm-will-move-away-from-silicon-at-7nm/
http://www.theregister.co.uk/2012/12/11/mit_non_silicon_transistor_tech/

[patrickjp93]

Yup, Germanium-Silicon fusion for 7nm and 5nm.

[RichardsD]

Well, hopefully the mass production of this will bring the cost down if Intel engineers decide this is the material for their next chips.

[ChrisxIxCross]

Yup, Germanium-Silicon fusion for 7nm and 5nm.

Wow really? I thought they couldn't go that low due to running into issues with quantum tunneling?

[patrickjp93]

Well, hopefully the mass production of this will bring the cost down if Intel engineers decide this is the material for their next chips.

PFFFT! Germanium is 4x more expensive to produce in the purity you need for a CMOS process.

[LiamApex]

i really kinda hope they stop using silicon as it will bring new things to the game new potentials for overclocking cant wait

[mr moose]

Sorry silicon, you had a good run, but now your ass is glass.

[LiamApex]

I HAVE IT, THE NEW MATERIAL INTEL SHOULD USE IS CUSTARD FOR THERE CPU 

[givingtnt]

pppft the last computer I built used redstone !

[mr.squishy]

I HAVE IT, THE NEW MATERIAL INTEL SHOULD USE IS CUSTARD FOR THERE CPU 

We can use Nutella for TIM (actually works btw)

[Aniallation]

already posted today

Please include a link to the original topic when claiming repost.

[handymanshandle]

Shouldn't making more cores, stronger cores be first priority then shrinking manufacture process?

[Norman_the_Owl]

Shouldn't making more cores, stronger cores be first priority then shrinking manufacture process?

No? Efficiency has been the name of the game for ever

[WelshDdraig]

Please include a link to the original topic when claiming repost.

The post I saw was talking about Moore's Law continuing through 7nm. Doesn't mention Intel abandoning Silicon. So I would say the article is fine.

Other Article ----> http://linustechtips.com/main/topic/315900-pc-world-intel-moores-law-will-continue-through-7nm-chips/ (Why can't people put links when saying "Repost"?)

[Guest Strangerbob]

wow... when we get to below 10nm, MAYBE there will finally be an actual point to upgrade a sandy bridge i5  

[SirSquid]

obviously intel is skipping 7nm for quantum computing processors   

[patrickjp93]

Wow really? I thought they couldn't go that low due to running into issues with quantum tunneling?

Hence the change in materials.

[patrickjp93]

wow... when we get to below 10nm, MAYBE there will finally be an actual point to upgrade a sandy bridge i5

There already would be a reason if software devs got off their lazy butts.

[Thony]

Excited. Probably gonna upgrade my cpu to one of these 7nm (2017/18).

[Deletive]

obviously intel is skipping 7nm for quantum computing processors

Quantum computing isn't going to be faster in the consumer market for a long long time. It's probably going to be slower for anything other then calculations (so most of the computer market) for awhile. It basically doesn't make anything faster it can just do things a considerably less number of operations.

Lets not forget all of the stuff that comes with it such as cooling it to Low kelvin temperatures due to Atom movement. (from what I understand)

I wouldn't even consider Quantum computing reasonable for another 15 years.

[LED_Guy]

Ignore all the comments about cost of SiGe (silicon-germanium) or InGaAs vs silicon. The cost of the materials is minor compared to the cost of processing. Fabrication costs (and more particularly yields) dominate CPU costs, not the price of the substrate.

[Opcode]

Not surprised with how much trouble they are having with just 10nm.

[patrickjp93]

Not surprised with how much trouble they are having with just 10nm.

10nm isn't giving them trouble. They haven't even finished building a 10nm fabrication plant. The one being built in Israel just began construction. Even sample runs won't begin until early 2016.

[LED_Guy]

Oh yeah, forgot to mention electron and hole mobilities in InGaAs are off the charts compared to silicon or even SiGe. III-V materials hit GHz frequencies a long time ago and can push past 100 GHz. Quantum confinement effects like 2-D electron gas devices offer all types of opportunities you can't get with group IV materials.

You also get the possibility of integrated optoelectronics so you can use optical signals instead of electrical signals.

Of course the trick is getting it all to work in devices with billions of transistors. . .

[Opcode]

10nm isn't giving them trouble. They haven't even finished building a 10nm fabrication plant. The one being built in Israel just began construction. Even sample runs won't begin until early 2016.

Cannonlake was postponed until 2H 2017 because of it. Parasitic capacitance was bad for 14nm and will be even worse for 10nm.