Friday, May 18th 2012

Samsung Electronics Presents a New Graphene Device Structure

Samsung Advanced Institute of Technology, the core R&D incubator for Samsung Electronics, has developed a new transistor structure utilizing graphene, touted as the "miracle material." As published online in the journal Science on Thursday, 17th May, this research is regarded to have brought us one step closer to the development of transistors that can overcome the limits of conventional silicon.

Currently, semiconductor devices consist of billions of silicon transistors. To increase the performance of semiconductors (the speed of devices), the options have to been to either reduce the size of individual transistors to shorten the traveling distance of electrons, or to use a material with higher electron mobility which allows for faster electron velocity. For the past 40 years, the industry has been increasing performance by reducing size. However, experts believe we are now nearing the potential limits of scaling down.
Image courtesy of Sammy Hub

Since graphene possesses electron mobility about 200 times greater than that of silicon, it has been considered a potential substitute. Although one issue with graphene is that, unlike conventional semiconducting materials, current cannot be switched off because it is semi-metallic. This has become the key issue in realizing graphene transistors. Both on and off flow of current is required in a transistor to represent "1" and "0" of digital signals. Previous solutions and research have tried to convert graphene into a semi-conductor. However, this radically decreased the mobility of graphene, leading to skepticism over the feasibility of graphene transistors.

By re-engineering the basic operating principles of digital switches, Samsung Advanced Institute of Technology has developed a device that can switch off the current in graphene without degrading its mobility. The demonstrated graphene-silicon Schottky barrier can switch current on or off by controlling the height of the barrier. The new device was named Barristor, after its barrier-controllable feature.

In addition, to expand the research into the possibility of logic device applications, the most basic logic gate (inverter) and logic circuits (half-adder) were fabricated, and basic operation (adding) was demonstrated.

Samsung Advanced Institute of Technology owns 9 major patents related to the structure and the operating method of the Graphene Barristor.

As demonstrated in this research, the institute has solved the most difficult problem in graphene device research and has opened the door to new directions for future studies. This breakthrough continues to keep Samsung Advanced Institute of Technology at the forefront of graphene-related industries.
Add your own comment

11 Comments on Samsung Electronics Presents a New Graphene Device Structure

#1
NC37
Intel says..."Woot!!! Ghz race is back baby!!! Take that ARM!!!"
AMD says..."At least our graphics still won't suck!!"
Posted on Reply
#2
Delta6326
This sounds awesome, to bad i could see the company's getting this tech, but only slowly raising the speed because say that you can take a 3GHz CPU(mid range)and now make it at 5GHz(still mid range of the new CPU's) most people in that product range now will not have to upgrade for years to come company's don't like that.

I honestly think lots of tech could be a lot faster now, but sense software programs are still playing catchup they aren't giving us their true power.

my crazy $0.02
Posted on Reply
#3
Frick
Fishfaced Nincompoop
by: Delta6326

I honestly think lots of tech could be a lot faster now, but sense software programs are still playing catchup they aren't giving us their true power.
I don't know, some software uses everything you throw at it.
Posted on Reply
#4
Vinska
by: Frick
I don't know, some software uses everything you throw at it.
More precisely: many programs keep finding new ways to waste any resources that our advancing technology can increasingly offer. [Averted in many games, though.]
In other words: one can usually get away with sloppy, easy-to-do and inefficient programming designs & using slow, but easy-to-use programming languages by throwing more processing power & memory into the equation. Also, knowing that such way to write programs is a great way to reduce development time, no wonder the world is almost universally falling into this temptation to program this way. :shadedshu
Posted on Reply
#5
phanbuey
True, but now look at the variety of software that you have and the applications that are available.

Sure my nokia was the fastest phone ive ever owned. But I can watch netflix on this one...

This is huge news - i remember when IBM started messing around with graphene and nanotech, and how promising that research was - this could really be a massive game changer in the way chips are manufactured in the near future.
Posted on Reply
#6
Vinska
by: phanbuey
True, but now look at the variety of software that you have and the applications that are available.
Yet, something always bugged me:
Most of the software I normally use is FOSS. When I am/was forced to run window$, quite a lot of software I had [or was forced] to use there was closed source/proprietary/non-free. Many FOSS programs I use are equivalents of those closed source programs, and not rarely, FOSS ones are more capable.
But the strange part of it was - FOSS programs are almost without exception have a ~5-20 times smaller [file]size.
Sometimes such difference is somewhat possible to explain, but when the difference in size becomes a couple of hundred megs or even over a gig... What. the. Butt!?

So, yeah, variety. But this 'variety' makes my brain stumble.
Posted on Reply
#7
phanbuey
by: Vinska
Yet, something always bugged me:
Most of the software I normally use is FOSS. When I am/was forced to run window$, quite a lot of software I had [or was forced] to use there was closed source/proprietary/non-free. Many FOSS programs I use are equivalents of those closed source programs, and not rarely, FOSS ones are more capable.
But the strange part of it was - FOSS programs are almost without exception have a ~5-20 times smaller [file]size.
Sometimes such difference is somewhat possible to explain, but when the difference in size becomes a couple of hundred megs or even over a gig... What. the. Butt!?

So, yeah, variety. But this 'variety' makes my brain stumble.
I rationalize it like this: Microsoft Excel. I use this all the time so I can tell the difference. I have open office too.

But MS excel has a TON more hidden features, programmed logic, prediction and text manipulation functionality, ability to program powerful macros, grab data off web pages... etc... So much so that people write entire applications in Excel. I had recently wrote an HL7 message spoofer: something normally done by a program like Mirth, that could log into our database using ODBC credentials, pull data, populate rows and create a bit-correct .HL7 file (with return carriage characters so it was recognized by the Unix HL7 engine), then write it to a monitored directory for the engine to pick up, with a unique ID.

Point is open office can't touch that. The complexity that allows that feature set to work is exponential.
Posted on Reply
#8
Vinska
@phanbuey

Hmm... point taken.

Yet, it does look wasteful, when an app consists of: >95% installer+uninstaller, <5% actual app (seen a few). Or ~40% installer+uninstaller, ~45% many layers of wrappers and some libraries for those wrappers, ~15% actual app (seen quite a lot).

P.S. Vinska hates anything related to office suites. And uses no more than Vinska is forced to. [lol, speaking in 3rd person, lol!]
Posted on Reply
#9
hardcore_gamer
by: Cristian_25H
The demonstrated graphene-silicon Schottky barrier can switch current on or off by controlling the height of the barrier
The transistor in graphene might have a structure similar to a MESFET. This is different from the MOSFET in silicon.

The gallium arsenide MESFET is able to operate at frequencies in excess of 250 GHz since the electron mobility of GaAs is much higher than silicon. This technology never took off because GaAs doesn't have a stable oxide layer like the silicon, and it was very difficult to scale MESFETs.

I hope graphene transistor has a different future.
Posted on Reply
#10
Darkrealms
Vinska you are looking at it with the wrong perspective. I can understand where you are coming from. I use both Linux and Windows, a more recent addition to the linux community. Much like phanbuey I love excel and it is wasted on 99% of the people that use it (I have open office as well). I also understand the simplicity of a well designed program in linux.
However..... The majority of the world is not us. It is full of people that would rather pay $$ than learn. Windows is sold complete, the programs install completely configured (majority of them), and the user can be dumb and make it work. There are many things in linux I still struggle with making it do, this is a training/learning problem on my end and I understand this. Most people would not and wouldn't bother trying...
Windows and their $$ software groups rule both the consumer and the enterprise markets because many (many not everyone) people are lazy or stupid and money is not scarce enough.
There are also some things paid for software does better.


Sorry on topic ; )
I'd love to see how this follows through, it would be a nice advancement. Especially for the mobile/compact markets.
Posted on Reply
#11
Vinska
I am wondering: (when|if) they improve this technology well enough to be able to make reasonable and practically usable integrated circuits... how well will they be in speed, component size and profitability? Silicon IC technologies will probably advance some more till that day comes. I bet when they reach "able to make graphene ICs for consumers, not just for it's R&D", it will probably be at least a little slower than the current IC, and much worse on component size (well, this is made mostly to allow making faster chips without reducing components), and most importantly - profitability. What I am saying: no one will mass produce chips that are unprofitable and cost a hell lot to produce (compared to what we have now). And that means - we will probably have to wait quite f***ing long till we will be able to say "My computer runs with a graphene CPU & GPU! Weee!!"
Posted on Reply
Add your own comment