The next step in nanotechnology | George Tulevski

Nearly every other year the transistors that power silicon computer chip shrink in size by half and double in performance, enabling our devices to become more mobile and accessible. But what happens when these components can't get any smaller? George Tulevski researches the unseen and untapped world of nanomaterials. His current work: developing chemical processes to compel billions of carbon nanotubes to assemble themselves into the patterns needed to build circuits, much the same way natural organisms build intricate, diverse and elegant structures. Could they hold the secret to the next generation of computing?

TEDTalks is a daily video podcast of the best talks and performances from the TED Conference, where the world's leading thinkers and doers give the talk of their lives in 18 minutes (or less). Look for talks on Technology, Entertainment and Design -- plus science, business, global issues, the arts and much more.

[RYSEproductions]

"Adding a handful of atoms"
Damn that's a lot of atoms

[LeoAtienza-tlum]

Please avoid using AI or even search your answer in google/internet.
Thankyou and Goodluck to your exam.

[unmedicateddepresso]

Kaway kaway sa mga nanonood dito dahil sa kinginang STS

[MrHansiping]

For those who got to the end and had no inkling what he's actually doing:

The challenge is very hard: You have a bunch of carbon nanotubes that you want to make into a computer circuit, you want to get them to sit down on a silicon wafer in a way that allows you to make billion-element circuits with switch sizes smaller than today's smallest transistors. How do you do this?

The solution, in principle, is actually very simple. You suspend a bunch of individual carbon nanotubes in solution. Then you make very fine patterns of things that stick nanotubes and don't stick nanotubes down on your silicon wafer. Then you just wash the nanotubes over the silicon wafer until they stick in the right place.

Well, in detail this is actually really really hard to achieve:

First, you need a way of generating patterns that are denser than those achievable by modern lithography. People have come up with very clever ways of doing this. Two of the best ways are something called block-copolymer lithography and DNA nanotechnology. In either case what you are doing is making chemical polymers that self-assemble together to create very fine patterns.

Second, you need a way of purifying semiconducting carbon nanotubes, cutting them to the exact right length for your circuits, then suspending them in solution. This is actually also no an easy problem, but people have come up with ways of making small polymers (eg, short strands of DNA) that bind onto carbon nanotubes, wrap around them, and suspend them in water.

Third, you need to create just the right interaction between the patterns on the surface and nanotubes in suspension for them to land on the right spots and wiggle into place along the patterns you define. This is pretty hard. You need to engineer the right chemical handles onto both the surface and the polymers wrapping the nanotubes. You might have to have the surface help channel stuff into the right place with morphological features. Then you create just the right conditions so that nanotubes will align themselves onto surface patterns via many weak chemical interactions (if the interactions are too strong they can lay down in the wrong directions, make tangled messes, etc). This is pretty hard. People are still figuring this out.

Lastly, you need to scale up all of these processes so that IBM or Intel can make perfect wafers with literally trillions of these devices. This is -really- hard because once you get down to nanoscales, thermodynamics fights against you every step of the way. You can make things easier by simplifying the requirements for assembly: eg instead of having nanotubes lay down on the surface to make circuits, just have them all forming parallel arrays pointing in one direction. Use lithography to define the other features that make up the circuit on top of them. Even so, it takes tons and tons of work.

Finally, here's the one question that nobody has asked: Why carbon nanotubes?

It turns out, it's not impossible for Intel or IBM to make transistors that are about as small as carbon nanotube transistors. The problem is that if the electricity flowing through the transistor is carried by silicon, there will be so much power dissipation that your CPU will literally melt itself before you can finish one game of Call of Duty. Semiconducting carbon nanotubes, are just about, the most power efficient semiconductors known. Graphene could be even better but it's actually not a semiconductor, so, no real good way to make it into a good transistor. So, carbon nanotubes are really, one of just a handful of materials that could possibly allow Moore's Law to continue down to transistors on the size of 10 nm wide or so, which is why IBM is still at it.

(BTW, 10 nm transistor is not the same as a 10 nm node on the semiconductor roadmap. The 10 nm node actually has much bigger transistors).

Anyways, I hope IBM succeeds. I'm rooting for them.

[neodark]

That's the kind of title Ted talks should have.

[Oddragnar]

tl;dr use chemistry to make carbon nanotubes assemble themselves to create a new generation of electronics, computers, clean tech etc

[MrMilwaukee]

"Elevator to space"

The Sun: that's illegal

[mindymurmur]

Hello 1st yr BSBA students from SFC nga tan aw ani because os STS.

[ultravidz]

Really bro? All that buildup just to tell us that we're missing "chemistry"? You're not even gonna elaborate on anything specific, like a new chemical process being developed or whatever?

[jbeegs]

An great example of an excellent science communicator!

[Endisupertramp]

New proposals for renaming TED:

VaS (Vague and Superficial)
SoNR (Science, only not really)
StS (Skimming the Surface)
BatBS (Beating around the Bush, Scientifically)
WLaS ( Waxing Lyrically about Science)

This guy's motto: Rome wasn't built in a day, neither will I get to the point in one.

TED's unnoficial mottos:

- Speak in multitudes but say nothing

- Science Lite

[rejoyy]

This guy so reminds me of Jeff Goldblum. He got the looks, speech, body language and even the glasses down pat.

[siegfredperez]

STS be like.
Watch the video on Youtube entitled "The Next Step in Nanotechnology" by George Tule ski, Ted Talks.

[billhopen]

As a sculptor, I gotta tell you you have to re-work your opening metaphore. "millions of tiny stone dust particles" can indeed be assembled, dude, its called clay, and you build it up by adding, building the form up in space, as opposed to the subtractive or carving method employed by a stone carver, removing dust.

[JesusChristDenton_]

We've been throwing the word "nanotechnology" around for decades yet, despite only our bests effort, we are only inching closer to that molecular-scale frontier when in fact we should be racing towards it. -Deus Ex Human Revolution

[thisbishawesome]

For anyone disappointed in the science depth or lack thereof the Ted talk. It's because he was basically a sales man trying to get investors/donors or whatever scientists call their source of money

[acegabrielcruz]

This shows how brilliant nano technology really is thanks STS

[Pakanahymni]

I remember nano-everything being on every single science mag cover 15 years ago.

[bv]

Wow, ideas actually worth sharing for a change! Stick with science, TED!

[veemacks]

Still damn waiting for a decent nano-coating I can have applied to my car (incl. windshield) so I never have to clean it again.