The graphene-based RF transistors are supposed to defeat the fastest transistors using the gallium arsenic method, and pave the way for fast and commercial carbon-based electronics.
The graphene-based RF transistors are supposed to beat the fastest gallium arsenide transistors in their speed, and pave the way for fast and commercial carbon-based electronics.
"Extraordinary claims have already been made regarding the ability of graphene for day-to-day use in transistors, but this is the first demonstration of a graphene-based RF transistor built under the relevant technological conditions and at the right scale," says IBM Fellow Pandon Avoris, who is responsible for carbon-based materials development efforts at IBM's Yorktown Heights labs. New York.
The graphene RF transistors were created for the Defense Department's Advanced Projects Agency under the Carbon Electronics for RF Applications (CERA) program.
The transmission rate is 4 times faster than the previous demos. The transistors were created on the wafer using a graphene growth process that is compatible with the process used for silicon transistors. CERA plans to incorporate the graphene transistors instead of the gallium arsenide transistors currently used in military communication systems.
The graphene wafers are created by taking wafers of commercially available silicon carbide (SiC), then burning off the top layer of silicon in a process known as thermal decomposition. The result is a single layer of graphene on the surface that would otherwise only be insulated. After that, the patterns of the graphene transistors were embedded using a metal top-gate architecture and then the graphene layer was isolated in the required places by a polymer.
The graphene transistor has 2 times stronger performance than silicon with the same gate length (100 GHz in graphene versus 40 GHz in silicon).
The gate width was 240 nm, 10 times larger than the smallest gate possible with existing lithography technologies (less than 35 nm). By optimizing the process to increase throughput and reduce gate length, IBM intends to increase the speed of graphene transistors up to 1 terahertz, the goal of the CERA program.
16 תגובות
Moish-
This transistor as intended for RF applications and not for computers.
It's too big
The fact that the gateway operates at 100 GHz does not mean that the computer will run at 100 GHz
Each operation in the set of operations of the computer is built from several operations of gates
To the editor: A little proofreading wouldn't hurt.
In the third paragraph: go out and don't go out.
Human proofreading can be done or the rose's spell checker can be used.
And if I can spot mistakes then anyone can.
Thanks Bully. fixed
First time I enter this site. And this is the first article I read here. I was disappointed by the amount of typos here. I would have expected a higher level.
Examples:
Transistors
Tranizatori
Silicone
extraordinary
The length of your hair
To akexd9 also: No - a graphene transistor of the same size as silicon or germanium silicon is much faster.
The reason we used a technology called hetero-junction/structure and in Hebrew a mixed structure.
In Intel chips, the base logic is also used in this way in my opinion (I work for Intel), but in silicon-germanium the performance is limited for reasons we will not specify. A mixed structure can be compared to 2 slices of bread with chocolate spread in the middle. The atoms that contaminate the conduction electron donors are inserted into the slices of bread. The energy level in the chocolate spread is lower and therefore the conduction electrons move there. If you place a battery perpendicular to the page, that is, the electron moves perpendicular to the page along the spreader - since there are no attracting contaminating ions, the average distance it accelerates is much greater than a normal semiconductor material - where the accelerating electrons collide with the contaminating ions from which they were donated and stop. The classical explanation - the semi-classical or quantum calculation.
As a result, the average speed of the electrons is even 100 times higher than their speed in a semiconductor made of one material: silicon for example. Electricians measure this in a parameter called the conductivity coefficient.
If the transistor crossing speed is higher, the frequency at which information can be changed is correspondingly higher. Therefore in the material
Like GAAS, a transistor channel width of tens of microns reaches frequencies of hundreds of GHZ, while in a silicon transistor with fractions of a micron, only 3GHZ.
At the time, there was an article on the website about mixed buildings - there, too, the intuitive explanation was reasonable but could be improved.
In addition to Eran M's correct words
The importance of the experiment that until today the world has been divided into two: silicon chips for computers (with some Germanium) - a cheap material, and Gallium Arsenide chips - expensive materials for selected RF and electro-optics applications. In silicon they have been shuffling with the frequency at 3GHZ for years, while in GaAs at 150GHZ for 10-15 years.
Here, for the first time, a window was opened to accelerate the processing speed in conventional chips. A problem that arises and blocks further miniaturization in silicon - the power increases as the square of the operating frequency. Therefore, a silicon chip will burn at 100MHZ and if you want to operate it at 4GHZ, cool it with nitrogen. In fact, the neurons in our brains are made of carbon-based materials and the amount of parallel processing in them is amazing. We don't burn out. It's possible in graphene - the power doesn't burn out the chip.
It is also clear that there is no problem of printing integrated circuits like GAAS and the power may be solved as well.
If the power problem is not solved, when it is clear that 1) the material cost problem has been solved here, 2) the ability to miniaturize compared to GAAS - they will continue to work on parallel processing of hundreds of processing units at the same time.
Is it graphene in the sense of graphene or graphene in the sense of graphene?
To Alex 7:
The answer to most of your questions: not clear, and even the researchers will have trouble giving you good answers.
In general, from a theoretical point of view, there are experiments, and some of them have even been published on this site, that theoretically it is possible to reach speeds of up to 100 from today, by using graphene. Also remember that standard Intel processors work at a few gigahertz.
Not so understood article: what is the importance of 100 gigahertz? Are the graphene-based transistors 2 times faster than existing transistors and this is a theoretical limit or can they be even faster? Does this mean that if we manage to create a processor from 35 nm transistors based on graphene, they will be exactly twice as fast as existing processors with current 35 nm technology?
What about the reduction of transistors based on graphene, who is involved in the issue and what are the expectations when it will happen if at all?
Finally a significant technological innovation for the opening of the millennium.
Until I was enthusiastic about the fact that Intel wants to produce in 11 nm, now we are talking about it?
What wonder we became indifferent to innovations 🙂
Ahh, so this is a recipe for sandwiches
Thank you Yigal,
I told myself that I would get to the end of the article and ask the meaning of the word.
You saved me.
Thanks Legal (1). I began to wonder what he meant
The poet Blizovsky.
Wafer = slice (Hebrew-Hebrew!)