The advent of atomic-scale transistors CPU operation is expected to break the bottlen

[myaomar1]

Hi evreone i read this article sow ammazing

University of Texas scientists have created the world's first silicon transistor ene, the thickness of only one atom. With this transistor, future scientists may someday produce less energy, power, faster chips.

　　With hot graphene, another similar nanometer new materials also aroused great interest academic material, which is silicon-ene (Silicene). This material may become a new cornerstone of future IT technology. This transistor has a huge potential for computer scientists Silylene is a more interesting than graphene materials.

　　This invention is contrary to many observers by surprise,
you can read more here : blog url removed

Because there is discussion and a better link below, I am leaving the thread here.
admin allisolm

 

[DrMrLordX]

Hmm. Some reason why you're linking to a blogspot page instead of a source article from a reputable source?

Like the one you quoted in your post? :

http://www.utexas.edu/news/2015/02/03/silicon-silicene-transistors/

Just curious.

 

[witeken]

They've made a single transistor, and I've read reports that silicene is suicidal, so we'll have to see if it ever gets used in chips. The next steps will be multiple transistors, simple logic circuits, they have to be the same size or smaller than current ones and have many billions of transistors on a chip, it needs to be able to be build in huge quantities, it has to have a good cost and yield, etc.
IIRC, there are already graphene transistors, maybe even some simple logic, so graphene really is more promising, at least with regards to TTM.

It's a tall set of requirements, and there are many competing technologies.

As of now, the roadmap looks ~~something like this for Intel:

10nm -- III-V and Ge
5nm -- GAA
3nm -- wires/dots?
2.5nm -- CNTs
1.2nm ~graphene
0.6nm ~spintronics
Beyond: vacuum transistor, quantum tunneling transistor, etc.

Please take "roadmap" with a grain of sand .

 

[Ken g6]

While we're discussing hypothetically-useful materials, let me introduce you to stanene. It's a single-atom-thick layer of tin, and in theory it may be a room-temperature "superconductor"! (Not a "real" superconductor, but close enough for electronics work.) I don't know if it can form transistors, but (if it works and can be built on silicon) it can certainly improve the wires between them and make chips run cooler.

 

[Idontcare]

They've made a single transistor, and I've read reports that silicene is suicidal, so we'll have to see if it ever gets used in chips. The next steps will be multiple transistors, simple logic circuits, they have to be the same size or smaller than current ones and have many billions of transistors on a chip, it needs to be able to be build in huge quantities, it has to have a good cost and yield, etc.
IIRC, there are already graphene transistors, maybe even some simple logic, so graphene really is more promising, at least with regards to TTM.

It's a tall set of requirements, and there are many competing technologies.

As of now, the roadmap looks ~~something like this for Intel:

10nm -- III-V and Ge
5nm -- GAA
3nm -- wires/dots?
2.5nm -- CNTs
1.2nm ~graphene
0.6nm ~spintronics
Beyond: vacuum transistor, quantum tunneling transistor, etc.

Please take "roadmap" with a grain of sand .

MoS2?

 

[witeken]

MoS2?

http://spectrum.ieee.org/nanoclast/...laces-silicon-in-the-transistor-of-the-future

Yes; it wasn't an exhaustive list.

 

[DrMrLordX]

I've read reports that silicene is suicidal, so we'll have to see if it ever gets used in chips.

The U of Texas article above covers this aspect of silicene as well as some of the work that has gone into isolating it from unwanted oxidation:

Despite its promise for commercial adaptation, silicene has proved extremely difficult to create and work with because of its complexity and instability when exposed to air.

To work around these issues, Akinwande teamed with Alessandro Molle at the Institute for Microelectronics and Microsystems in Agrate Brianza, Italy, to develop a new method for fabricating the silicene that reduces its exposure to air. To start, the researchers let a hot vapor of silicon atoms condense onto a crystalline block of silver in a vacuum chamber. They then formed a silicene sheet on a thin layer of silver and added a nanometer-thick layer of alumina on top. Because of these protective layers, the team could safely peel it of its base and transfer it silver-side-up to an oxidized-silicon substrate. They were then able to gently scrape some of the silver to leave behind two islands of metal as electrodes, with a strip of silicene between them.

Sounds like a lab-only solution right now, as is so often the case.

 

[Idontcare]

The U of Texas article above covers this aspect of silicene as well as some of the work that has gone into isolating it from unwanted oxidation:

Sounds like a lab-only solution right now, as is so often the case.

Actually the unwanted oxidation was/is truly just a gate-stopper for producing proof-of-concept devices from which electrical characteristics can be measured.

I.e. it is a barrier to being able to build the prototypes, but wouldn't be a barrier were the device to make it into an R&D line or production fab.

But it will never make it into an R&D line or production fab if the prototypes aren't built in a uni-lab so people can determine if it is the right path to invest billions of R&D into. (path finding)

So no worries about the trials and tribulations of developing one-off lab-only solutions needed to enable the creation of a few electrically functioning prototypes.

Those challenges aren't the ones that translate into the fab environment. When/if that time comes then the engineers will have the right tools and machines to get the job done, entirely in vacuum if necessary.

 

[oobydoobydoo]

Why would oxidation be such a big obstacle? You would think they would have the ability to build a device that produces a hard vacuum and then stick the device that makes these silicene chips inside that hard vacuum.

 

[Xpage]

or just work in a non oxidizing, N2 environment.

 

[DrMrLordX]

Actually the unwanted oxidation was/is truly just a gate-stopper for producing proof-of-concept devices from which electrical characteristics can be measured.

I.e. it is a barrier to being able to build the prototypes, but wouldn't be a barrier were the device to make it into an R&D line or production fab.

But it will never make it into an R&D line or production fab if the prototypes aren't built in a uni-lab so people can determine if it is the right path to invest billions of R&D into. (path finding)

So no worries about the trials and tribulations of developing one-off lab-only solutions needed to enable the creation of a few electrically functioning prototypes.

Those challenges aren't the ones that translate into the fab environment. When/if that time comes then the engineers will have the right tools and machines to get the job done, entirely in vacuum if necessary.

Makes sense, especially when you consider what must be done to fab modern chips in existing fabs.

Why would oxidation be such a big obstacle? You would think they would have the ability to build a device that produces a hard vacuum and then stick the device that makes these silicene chips inside that hard vacuum.

Just up-front costs, which are already high for fab equipment anyway. The interesting part isn't what they do to produce the silicene transistors, but what they (currently) do to protect them once the process is complete.

or just work in a non oxidizing, N2 environment.

N2 might actually react with the Si vapor, though I am too l4m3 to do any Gibbs equations (or what have you) to examine the possibility of that happening. Still, silicon doesn't boil until 2357C (1 atm I guess). Lots of stuff reacts at that temperature.

 

[ctsoth]

+1 vote for argon, yet I admit I don't know if even a trace of oxygen, lets say 1-10ppm is enough to cause problems for this process

 

[DrMrLordX]

Argon should be safe. Regardless:

http://en.wikipedia.org/wiki/Silicon_nitride

3 Si + 2 N2 -> Si3N4 @ ~1300C

So yeah N2 is probably out.

 

[svarog19]

It is a waste of time to just search for better materials, main issue is that there are really no new CPU micro architectures, we still have POWER, ARM, x86, etc... x86 is restrained by maintaining legacy code and POWER and ARM are restrained to maintain similarity with previous generations respectively.

MIPS is practically dead from what I know and because of that we don't know how well it would perform if it had same amount of development as ARM, POWER or x86.

I wonder if MILL architecture will ever be picked up for sake for workstations for rendering and such even though its design is terrible for yields because how its works and it would be nearly the same as IBM producing POWER CPU's with terrible yield for sake of performance of CPU's that survive....

 

[Headfoot]

It is a waste of time to just search for better materials, main issue is that there are really no new CPU micro architectures, we still have POWER, ARM, x86, etc... x86 is restrained by maintaining legacy code and POWER and ARM are restrained to maintain similarity with previous generations respectively.

MIPS is practically dead from what I know and because of that we don't know how well it would perform if it had same amount of development as ARM, POWER or x86.

I wonder if MILL architecture will ever be picked up for sake for workstations for rendering and such even though its design is terrible for yields because how its works and it would be nearly the same as IBM producing POWER CPU's with terrible yield for sake of performance of CPU's that survive....

Let's take a trip down the derail trail