IBM Shows 155GHz Graphene Transistor

Someone enlighten me here. I thought that transistors were originally analog devices for many years. They had to be re-designed to be on-off devices. Am I wrong?

You're right, today's transistors are like a faucet that is used only at full close or full open but still have many intermediate positions

I thought the transistor was suppose to be an on/off switch? I congratulate the engineers at IBM, yet the transistor itself is not all that complete.

Can we communicate with Aliens on their radio frequencies now?

I wonder how small these things are. Graphene is a single-layer sheet of carbon like the article says, but you can stack layers on top of each other like a pile of paper. Is the transistor based on just a single layer, and if so, what surface area is required? Even a transistor involving 10 hexagons in each direction would only have dimensions of 2.46 nm on a side (one carbon bound is 1.42 Angstroms, and there are sqrt(3) bond-lengths per hexagon).
Unfortunately it sounds like these are more akin to bipolar transistors than cmos in terms of their electrical characteristics, so the amount of current required to turn and keep one "on" rather than "off" would be prohibitive on a chip-wide scale. But if I wanted a radio with a 3mm wavelength, these could be used in the amplifier circuit if the gain is high enough.

one comment

The advantages of graphene do not come from transistor size but the fact that it's a superconductor at normal temps. Because it's a much better conductor anything built with it produces much less heat. This allows them to get to these high frequencies.

A transistor is used as an ON/OFF switch in digital application. But it's used as an amplifier in analog application.

The transistor can be in the active region, saturation region and cut-off region. So, it is an analog device, but in the saturation region is fully on, while in the cut-off region is off. These regions are used for the "digital" signal.
I guess most of you guys know only how to play Cryis. :-(

In IE9, comments don't work. The button does nothing. One has to use compatibility view in order to post a comment. Come on guys, it has been more than a month, now.

We are not ALL electrical engineers afterall haha

For high power analog applications, it is necessary to use a feedback temperature sensor on transistors to compensate for how the temperature of the device affects the active region. Hopefully, the graphene transistor will be scaled for these types of applications and be less susceptible to drift of the active region. Oh, and of course I would like to see them used in my next CPU, as well.

Graphene Transistor has zero bandgap, which is another form of a fancy resistor?

Then take 5 minutes to scan wikipedia before making a comment. Not only will you sound vastly more intelligent, you'll also manage to learn something today, as well as contributing something useful to the comment section.

And before this thought is bashed, I'm not saying wikipedia will make you, or make you sound more intelligent. The active pursuit of knowledge will.

Time to clear up some misconceptions here:

First, all transistors can be operated as analog devices or digital devices. Obviously, many are optimized/intended for use as one type or the other, but almost any transistor can be used in either mode.

Second, graphene is nowhere near a "superconductor" at normal temps. Not even close, and that's not the point of using graphene for anything. A true room-temperature superconductor will change the world in ways this transistor will not.

What is true, however, is that this is pretty a cool device.

Transistors are used for modulating voltage or current in a circuit based on an input. They can be thought of just like a faucet. Your input on the handle can make the water run slow or fast or not at all.

Graphene is not a superconductor. It would be great if it were, but it isn't (at least not anything remotely resembling normal temperatures).

As far as I know, MOSFETs can be manipulated into electrical switches.

Yup we can play Crisis, but could a computer built with these transistors ;-)

looks like chicken wire

Just to correct some severe misinformation from zzz_b:

Transistors have three modes of operation: Cutoff, linear (aka triode), and saturation (aka forward active).

Analog circuits generally keep transistors in the saturation region.

Digital circuits generally switch transistors between cutoff (off) and linear (on).

Wrong: Analog circuits generally use the linear, and maybe cutoff regions. Digital circuits use the cutoff and saturation regions.

Also, why is anyone impressed by a single transistor operating at 155 GHz??? There have been silicon transistors that operate faster! http://www.betanews.com/article/50 [...] 1150919128 350 GHz at room temperature, 500 GHz near 0 K.

There are perfectly formed crystals grown in the space station that have oscillated at 3-4 THz. I don't think speed is the only metric here. The fact that these are such a new medium offers many niche applications that are not possible with standard silicon mediums.

This is undergrad-level stuff, and you're clearly a bit out of your area of expertise. I think you have just enough rudimentary understanding to cause you trouble, because you're aware that analog amplifiers have linear input-output response, whereas any digital circuit is highly nonlinear. You're confused about what the linear and saturated regions actually mean, and you're confused by the different names used by bipolar and MOS devices.

Here's your basic course in devices:

Bipolar devices are used in the saturated region for digital circuits, so you accidentally got that one right. But no one has seriously used bipolar for digital since, oh, 1985 or so.

Bipolar devices are used in the forward active region for analog.

MOS devices are used in the triode/linear region for digital (this is the same high-slope I-V region which is unfortunately called "saturated" for bipolar devices). Here's your special warning: This does not mean that digital circuits are linear. Look at an I-V curve for a MOS device sometime.

MOS devices are used in the saturated region for analog. This is the same region called "forward active" for bipolar devices.

If you really want, I can break out some text books and quote you chapter and verse. But I'd prefer not to, considering that you could search this information on your own faster.

Shame, a 155GHz CPU was looking like a lot of fun till I read a bit more.

But this is IBM we are talking about, give it 10 years and there will be graphene CPUs belting out 300GHz of pixel-pounding-power in the Xbox-4

...aah, perfect for building my polychotomous state variable computer.

OK, so they use nearly opposite terms for MOS and bipolar. I am familiar with the workings of both, but not with the MOS terminology. I'm not an expert, but it is kind of a hobby. Thankyou for correcting me.

Yeah, seriously. The terminology is pretty unfortunate.

Btw, I also prefer Linux, but you can school me on that in another thread.

Wow, arrogant nerds? I'm shocked.

Take your knowledge of EE and stick it up your ass. Nobody gives a fuck you know everything about transistors, stop using it to try and make yourself feel superior to other people.