IBM Scientists Build Computer Chips From DNA

snydeq writes "Scientists at IBM are experimenting with using DNA molecules as a way to create tiny circuits that could form the basis of smaller, more powerful computer chips. The technique builds on work done by Cal Tech's Paul Rothemund, who found that DNA molecules can be made to 'self-assemble' into tiny forms [PDF] such as triangles, squares and stars. 'To make a chip, the scientists first create lithographic templates using traditional chip making techniques. After, they pour a DNA solution over the surface of the silicon and the tiny triangles and squares — what the scientists call DNA origami — line themselves up to the patterns etched out using lithography.' DNA-based chips may sound like crackpot tech, but those involved believe the methodology could lead to a new way of fabricating features on the surface of chips that allows semiconductors to be made even smaller, faster and more power-efficient than they are today."

Comments

Nanofabrication

[neonprimetime]

"The degree of difficulty of nanofabrication is going up rapidly," Wallraff said.

Tell me about it! I can hardly keep up.

And when will it become self-aware?

[sxltrex]

It's only a matter of time. We're screwed. Nice going, IBM.

Re:And when will it become self-aware?

[garyisabusyguy]

I, for one, look forward to welcoming our bio-nano-tech overlords

Re:And when will it become self-aware?

[aphelion_rock]

More to the point, when will it become self reproducing? Add to this the ability to reproduce and mankind becomes god. Do we then put it on another planet and see if it survives and evolves...

Re:And when will it become self-aware?

[falconwolf]

I, for one, look forward to welcoming our bio-nano-tech overlords

I doubt it will happen in my lifetime but I'd like to see when either nanotechnology or Neurogenesis can repair damaged brains. I survived a Traumatic Brain Injury, TBI, and I'd be in line as a test subject.

Either that, or be in line to transplant one of Marvin's brains.

Falcon

Re:And when will it become self-aware?

[PitaBred]

DNA is just information. How well does a hard drive compute? Not very well. It has some computing capability, but to really make it do something interesting you have to hook it up to the rest of the computer. Same with DNA in relation to cells.

Re:And when will it become self-aware?

[interkin3tic]

DNA seems to have been in use for 4 billion years. Humans arose 200k years ago, some humans can be described as "self-aware."

So we have at most 3.8 billion years to decide if we are going to fight or welcome our self-aware DNA computer overlords.

Re:And when will it become self-aware?

[FudRucker]

that is a slippery slope, next thing you know they will be wanting equal rights and vote...

Re:And when will it become self-aware?

[Anonymous Coward]

Yeah and then they'll vote in a block and they'll elect the teleprompter as the next president. Where will it all end?

Re:And when will it become self-aware?

[superwiz]

that's a false analogy. it presumes that design that is intelligent (by the virtue of being done by humans) is as inefficient as the design based on (often random) restrictions posed by natural conditions and natural adaptation to those conditions renewed by an occasional random mutation.

Re:And when will it become self-aware?

[interkin3tic]

The chips aren't self-aware now, or even really thinking on their own presumably, and they aren't designed to do so. To go from unthinking chips to self awareness on the level of humans would likely be more inefficient and unlikely as our evolution has been. Also given restrictions on the system, it's more unlikely that the computer would continue to function as the thing started changing, not to mention it would have to evolve with much fewer resources in a much more closed space. So yeah, I think it's a false analogy: the computer evolving intelligence is far less likely and would take a lot longer probably.

Also I was being facetious.

Re:And when will it become self-aware?

[superwiz]

It's a false analogy because this is a mechanism for designing intelligent being from scratch. My point was that just as in everything else, man-made things will be better than natural things and they will get designed much, much faster. The technology allows for arranging DNA in a pre-designed sequence. Which is the first step to treating creation of new organisms as a programming task. No, that's not what the chips are designed to do, but it is the side-effect of having this technology available now.

Re:And when will it become self-aware?

[davester666]

But, you will have to worry about whether these nano-bots will simply stop after restoring the brain to it's previous state (so I guess you would forget what happened to you since before the 'incident', or will the bot's become self-aware (possibly as a group), then rewire your brain so your body is totally controlled by the bot's...

Re:And when will it become self-aware?

[falconwolf]

But, you will have to worry about whether these nano-bots will simply stop after restoring the brain to it's previous state (so I guess you would forget what happened to you since before the 'incident', or will the bot's become self-aware (possibly as a group), then rewire your brain so your body is totally controlled by the bot's...

No, I don't worry about that. I think if anything it'd be a relief for me.

Falcon

DNA is smaller?

[Anonymous Coward]

How do they intend to make circuits smaller than current lithography methods allow when the DNA solution is aligning itself to lithograph etchings? Further work required, methinks.

Re:DNA is smaller?

[Chris Burke]

DNA molecules are put inside the lithographic channels, therein form structures which are smaller than the lithography itself. Seems straightforward to me... It's just a way to guide their self-organization. Think of it this way -- the tiniest lithographic mark you can make can be used to make one transistor gate. Or, in the same space, you can get the full complexity of one or more DNA molecules.

Re:DNA is smaller?

[mgblst]

How do they intend to make circuits smaller than current lithography methods allow when the DNA solution is aligning itself to lithograph etchings? Further work required, methinks.

No, they have finished. It is ready. You can get it on at your local best buy right now. Just go and ask for your DNA laptop, they will now what you mean.

Japanese Words

[JKWSN]

pour a DNA solution over the surface of the silicon and the tiny triangles and squares â" what the scientists call DNA origami

I think I saw something like this on the internet once but a different Japanese word was used...

I know first "hand" this doesn't work

[Anonymous Coward]

After, they pour a DNA solution over the surface of the silicon and the tiny triangles and squares line themselves up to the patterns etched out using lithography"

I've been pouring a DNA solution over the surface of my computer for years. Doesn't do anything but break the keyboard. This story is bunk.

Lithography-based sounds like a good idea

[Chris Burke]

I remember over a decade ago hearing about research where they had demonstrated DNA-based computation. They'd used the molecules to perform some reasonably complex algorithm and got the correct answer. It was extremely fast, in part due to using a lot of parallelism. The only problem -- the 'answer' was somewhere in the beaker full of DNA goop and had to be chemically sorted out to actually see what happened. So, uh, not terribly practical.

Using lithography to put molecules where you want them to be sounds a lot better than a beaker of goop. :)

Re:Lithography-based sounds like a good idea

[CannonballHead]

Hm... but just imagine what a beowulf cluster of cloud-based goop computing systems could do?

it sounds like a +5 Funny Mod Comment at any rate. ;)

Re:Lithography-based sounds like a good idea

[Entropic Alchemist]

In 1994 Leonard M. Adleman solved the travelling salesman problem in 7 days with a DNA computer. I think that is probably what you are thinking of. (Article about it here)

Re:Lithography-based sounds like a good idea

[Anonymous Coward]

Yes, and by the way, Len Adelman is the A in RSA.

Re:Lithography-based sounds like a good idea

[Chris Burke]

Yes, that was it! I remember it was the traveling salesman problem. Pretty impressive for a brand new method of computation.

Does it self-replicate Linux?

[erroneus]

To be "DNA" doesn't it need to be able to self-replicate or something like that? Given the appropriate raw materials, will the DNA chips self-replicate and expand themselves?

Re:Does it self-replicate Linux?

[DragonWriter]

To be "DNA" doesn't it need to be able to self-replicate or something like that?

No, to be "life" it needs to do that (among other things.) To be DNA, all it has to do is, well, be DNA.

Re:Does it self-replicate Linux?

[Anonymous Coward]

So you're on what exactly? Ultra-meth? Super-crack? Do you even know what DNA is?

Re:Does it self-replicate Linux?

[Anonymous Coward]

Fuck you nigsausage!

Port the process to RNA...

[swanzilla]

I foresee a bunch of very angry and utterly confused "life starts at DNA" people.

Re:Port the process to RNA...

[VampireByte]

Especially if stem cells get added into the mix.

Re:Port the process to RNA...

[Tubal-Cain]

Why?

Re:Port the process to RNA...

[falconwolf]

I foresee a bunch of very angry and utterly confused "life starts at DNA" people.

If "life" means being able to reproduce then because RNA can self replicate the quote needs to be "life starts at DNA or RNA".

Falcon

No semen joke yet?

[Anonymous Coward]

*Pinching myself*

Re:No semen joke yet?

[Anonymous Coward]

*Pinching myself*

Is that what the kids are calling it these days? Or is that just because it's that small that that's the only way to stimulate?

Re:No semen joke yet?

[speedingant]

lol

Re:No semen joke yet?

[Anonymous Coward]

lol

You missed it http://hardware.slashdot.org/comments.pl?sid=1338969&cid=29098355

Good for the environment?

[LuxMaker]

Feed it beans and watch it output vapor....ware.

why DNA?

[panthroman]

I'm trying to figure out why the researchers are using DNA. Is it...

A - some unique and intrinsic property of DNA that makes it suited for the job.
(If so, then is it just coincidence that our genetic information is stored in a molecule that has these unique properties?)

B - just that DNA has been so well-studied in the last half century that we can manipulate it better (and cheaper) than most other complex micro-structures?
(If so, then that's just one more example of basic research leading to unforeseen breakthroughs.)

Anyone know which it is?

Re:why DNA?

[confused one]

self assembly and the ability to modify the structure of the molecule using simple methods to serve a specific purpose.

Re:why DNA?

[PotatoFarmer]

According to the article, it's a little from column A, and a little from column B - DNA has an intrinsic ability to self-assemble (A), and it's been studied to the point where the resultant forms of that self-assembly are pretty well known (B).

From reading TFA, it sounds like they're using a traditional lithographic technique to produce the substrate that is filled in by DNA. This DNA in turn self-assembles into structures that are more complex than can be reliably produced by lithography alone. Those structures are then coated in nanoparticles to form the actual IC interconnects.

Re:why DNA?

[Peter Steil]

More than likely because in the present age anything with even a hint at DNA or genetics tends to bring up stock value. Think of it like a buzz word.

Re:why DNA?

[ZankerH]

It's a well-studied and well-understood, large molecule, which is used for carrying information in nature and therefore well-suited for the task.

Re:why DNA?

[DragonWriter]

I'm trying to figure out why the researchers are using DNA. Is it...

A - some unique and intrinsic property of DNA that makes it suited for the job.
(If so, then is it just coincidence that our genetic information is stored in a molecule that has these unique properties?)

That's part of it; I wouldn't describe it as coincidence that a molecule that forms one of the basic information storage mechanisms which allows life to exist has properties that are useful in information technology, though.

B - just that DNA has been so well-studied in the last half century that we can manipulate it better (and cheaper) than most other complex micro-structures?

Well, yeah, there is some of that, too.

Its not Digital DNA...its DNA Digital!

[bezenek]

Todd

Nano-fabrication

[Thuktun]

I wonder if this is a practical step towards the ubiquitous "matter compilers" featured in Stephenson's The Diamond Age.

Re:Nano-fabrication

[cfa22]

Doubt it. But since the pieces of DNA used here are like stiff little rods with unique registry it's a bit closer to the wetware computation featured in Stephenson's The Diamond Age. Also because the DNA only does exactly what the designers tell it to do, it's a bit closer to the live-narrated children's storybook featured in Stephenson's The Diamond Age. Then again, such self-assembly might be useful in manufacturing those kick-ass skateboard wheels featured in Stephenson's The Diamond Age. Damn it, someday something in that book is going to come true.

Re:Nano-fabrication

[Anonymous Coward]

Probably the scene with the glow-in-the-dark condoms.

Macro processors?

[BlueKitties]

Instead of going small, why don't we go giant? My x86 processor is about an inch across, why not make it a foot? Do we really need *smaller,* why can't *bigger* be faster? Imagine our sleek Quantum Computing chip, and then blow it up macro size to a foot version. Why do we take a speed hit for not being microscopic? Why does being a "big processor" make it slower than a "small processor." Why can't we make jumbo versions of Quantum Computers? (I'm not familiar with logic gate circuitry, enlighten me please.)

Re:Macro processors?

[Peter Steil]

Several factors here: Cost longer distance = thicker interconnect required -> (see 1) Greater size = greater distance = longer travel time = slower Power consumption increases as component size increases whereas power consumption decreases as things get smaller.

Re:Macro processors?

[Peter Steil]

Several factors here:
Cost
longer distance = thicker interconnect required -> (see 1)
Greater size = greater distance = longer travel time = slower
Power consumption increases as component size increases whereas power consumption decreases as things get smaller.

Re:Macro processors?

[newcastlejon]

I'm not familiar with it either but I imagine that it becomes harder to maintain a synchronous clock signal (moot for designs that don't use one, natch) when you increase the size. It would also increase the amount of time that any signal takes to travel from one end to the other. As far as quantum goes, you're using individual or small groupings of atoms - that's just what size they are. The bulk of the machines I expect is made up of vacuum and cryogeny gear.

Re:Macro processors?

[Anonymous Coward]

AMD recently proved they can still make bigger processors. They are exactly as fast as normal/small processors, in frequency.
They are bigger, compute four times as much stuff in parallel, and consume four times as much power.

Power consumption is quadratic with size, if we suppose the processor is square, and linear with surface in general.

In comparison, if you take a Pentium II, build it with today's techniques (32nm etc) - and a little updated - you get an Atom. The Atom is way faster than a P2.
The reason is, the smaller transistors are, the faster they change their state. The speed of a processor is directly linked to the speed of each transistor. You get the equation...

BTW, it's relatively easy to build a massive ultra-violent computer. Just build a network of computers, and run some massively network-parallelized software. It's usually called a Supercomputer.

Re:Macro processors?

[lena_10326]

In addition to the other responses... there's also the reason of fabrication defects. The probability of a defect is constant for a 1cm x 1cm area so when the CPU area goes up the number of defects goes up. You only need 1 defect to trash the entire chip, so there's an financial need to achieve a defect rate per CPU as close to 0 as possible.

Re:Macro processors?

[Z80a]

besides the reasons posted down there, there is the yield factor too.

if i'm not mistaken is quite hard to make a chip that large that actually works, as you would need to do multiple precise "etchings" and they need to be EXACTLY lined up.

and also i heard that each silicon wafer costs around 2000 bucks,so if you get 4 chips working out of it, you will need to charge more than 500 dollars per chip.

Re:Macro processors?

[confused one]

above a certain size, it becomes harder to form defects significant enough to cause failure. I think around 200um you don't even need a clean room.

Re:Macro processors?

[mgblst]

I know, and why don't they make it a million meters wide, that would be great, so much more power. And they should make it red, that would be faster. And they should make it absorb heat, rather than produce heat, then we could sit it in the desert. And they should make self replicate, so every hour it reproduced itself, from the sand int the desert. Gee, if they just did that, we would have loads of cheap computing power...

Re:Macro processors?

[BlueKitties]

Imagine a miniature watch. Complete with miniature gears, and belts... and hands. But we can't build it, because it's too small. Oh well, we'll never ever have a watch, because it absolutely must be tiny tiny tiny to keep the time good.

International Borg Machines

[wa2flq]

Where's a Borg Queen when you need one?

It's not "Cal Tech"

[Anal Surprise]

Every time I read "Cal Tech", I cringe. "Caltech", please.

Re:It's not "Cal Tech"

[cfa22]

And we prefer "Cal" to "Berkeley." Also, in America, commas and periods go inside quotes.

Re:It's not "Cal Tech"

[Anonymous Coward]

Blatantly offtopic and nitpicky, but a pet peeve of mine: this is Slashdot, where most of us rather enjoy programming and other formal languages with an emphasis on syntax rules that make sense*. Punctuation is not part of what is being meta-referred to in this situation, so it makes no sense* to include it in the quotes. Natural language be damned.

* as defined by my personal and arbitrary idea of sense

Re:It's not "Cal Tech"

[Anonymous Coward]

And we prefer "Cal" to "Berkeley." Also, in America, commas and periods go inside quotes.

America? Which country?

Re:It's not "Cal Tech"

[Tubal-Cain]

Also, in America, commas and periods go inside quotes.

It's a completely unintuitive convention. The end of (or pause in) my sentence was not part of what the person I'm quoting said. Disregarding that rule is a conscious choice for some of us.

Re:It's not "Cal Tech"

[MindspanConsultants]

Style depends on where you live and what conventions you follow. In Canada, we seem to be adopting American spellings and punctuation rules I think more as a result of the auto-correction features and misspelling underlines in Firefox et al than a desire to conform to your standard. Laziness (failing to employ Canadian English dictionaries etc.) results in homogeneity favouring (yes there's a 'u' in there) the culture that designed the tool. Here's the breakdown on punctuation -- I happen to follow the British style myself.

Re:It's not "Cal Tech"

[Anonymous Coward]

there's a science term for this feeling on the internets, closely related to your nickname.

diden't we just see this on the last ep of eureka?

[Joe The Dragon]

diden't we just see this on the last ep of eureka?

self replication of DNA

[falconwolf]

To be "DNA" doesn't it need to be able to self-replicate or something like that? Given the appropriate raw materials, will the DNA chips self-replicate and expand themselves?

DNA can and does both self assemble and self replicate. I'm reading an article in "Tech review" on this subject, "First Life and Next Life". In experiments the author showed that the, artificial, DNA could also evolve.

Falcon

I for one welcome

[Anonymous Coward]

I for one welcome our new Terminator overlords that can travel through time without Arnold Schwarzenegger's face.

Identify a crackpot

[gr8_phk]

DNA-based chips may sound like crackpot tech, but those involved believe the methodology could lead to a new way of fabricating features on the surface of chips that allows semiconductors to be made even smaller, faster and more power-efficient than they are today.

You can tell if they're crackpots by weather they can complete the sentence "the methodology could lead to a new way of fabricating features on the surface of chips that allows semiconductors to be made even smaller, faster and more power-efficient than they are today BY DOING xxxxxxx yyyyyyy zzzzzzzz".

If they don't have a proposed method for reaching the goal, they're just blowing hot air. We see this all the time with people looking for publicity or money.

Who's DNA ?

[Kulfaangaren!]

My only question is...Who's DNA ? If it's Linus's I have no problem, but if it is Bill's or Steve's DNA I think we should burn the lab to the ground! Get your torches and pitchforks!?!?

Re:Who's DNA ?

[bobdehnhardt]

Richard Daystrom's, of course.

And you call yourself a nerd....

hook, line and sinker

[Goldsmith]

The PR people who have been pushing this research around the internet today did a truly amazing job. They took a pretty good surface science paper and made it look like the breakthrough of the year.

This research had nothing to do with computer chips. DNA origami has been around for a few years now, and this is the application of it to surfaces. When they have a *single* electrical measurement, you can start to get excited about electronics. When they have a functioning transistor, I could even live with a Slashdot headline about computer chips. This is jumping way ahead for a bunch of insulating organic molecules patterned on a surface.

It's alive, Jim

[Louis Savain]

This will give a new argument against those who object to the possibility of building truly intelligent computers on the basis that they are not living organisms.

Cool!

[Anonymous Coward]

Put 'em in my forehead or hand, and I can buy things with them!

Re:Cool!

[SilverHatHacker]

-1 Apocalyptic

Bio nano technology

[r45d15]

With bio-nano-technology you'll soon be able to replace your videocard by pulling the plug into your ass.

Caltech

[drerwk]

Not Cal.

When the technology gets perfected...

[Khyber]

Maybe they could call it the "Cell" processor. :P

Re:When the technology gets perfected...

[Anonymous Coward]

Until then it will be known as Codename: Scofield

What OS does it run

[Anonymous Coward]

Does it run Linux?

Re:What OS does it run

[Tubal-Cain]

I doubt it can even run "Hello World!" at this point.

DNA used to build machines

[Orion Blastar]

sounds like the Battlestar Galactica remake on SciFi ir SyFy network. Human looking Cylons.

That's the worst that can happen, right?

Guess

[Tubal-Cain]

How many countries have "America" in its name? Mexico is formally known as the "United Mexican States", and as far as I can tell Canada is just "Canada".

Not Borg or Skynet...

[jd2112]

This is Cylon ver. 0.1

when will it become self reproducing

[falconwolf]

DNA is already self replicable.

Do we then put it on another planet and see if it survives and evolves...

Meteorites, such as the Murchison meteorite have been found with organic matter. It contains amino acids as well as inorganic precursors that the Miller-Urey experiment proved could produce organic compounds.

Falcon

Re:when will it become self reproducing

[pinkushun]

Thanks for the links, very informative reads :)

Re:when will it become self reproducing

[Octogonal Raven]

DNA is already self replicable.

Do we then put it on another planet and see if it survives and evolves...

Meteorites, such as the Murchison meteorite have been found with organic matter. It contains amino acids as well as inorganic precursors that the Miller-Urey experiment proved could produce organic compounds.

Falcon

The problem is that the Miller-Urey experiment was irreparably flawed from the get-go. Note that they used water, methane, ammonia, and hydrogen. They used chemical compounds that would give them the results they so desperately needed to bolster their claims that life could have arisen on its own. The experiment, so flawed, only shows that blasting a bunch of water, methane, ammonia, and hydrogen with several thousand volts will create a few amino acids. (Five, to be exact, out of twenty-two required for life to exist.) The other flaw in the experiment, and indeed the evolutionary theory as a whole, is the law of entropy. An argument can be made that there is proof in computers or cars or what-all that it's possible to get a more advanced thing as time passes, contrary to said law. There is a common factor in all of those things though. Can you guess? Yup, a human mind. A guiding intelligence, if you will. In case you haven't noticed yet, I do not adhere to the so-called "theory of evolution".

Instead of going small, why don't we go giant?

[falconwolf]

I want to be able to carry it with me. It's a plus if I can have one that I control implanted.

Falcon

My only question is...Who's DNA ?

[falconwolf]

f it is Bill's or Steve's DNA I think we should burn the lab to the ground! Get your torches and pitchforks!?!?

Which Steve? Ballmer? Jobs? Or the Woz?

Falcon

Correction

[Anonymous Coward]

It's Caltech, not Cal Tech.

Codename

[Anonymous Coward]

Future IBM CPU Codename: Scofield

Not tonight darling, I've got headache...

[viraltus]

Never thought I might fix my computer with an aspirin.

Thanks for the links, very informative reads :)

[falconwolf]

If you found those interesting the article "First Life and Next Life" in the June issue of Tech Review should be more interesting. I just happened to be reading it when this thread started.

Falcon

Important nitpick

[hansraj]

Actually a particular instance of the traveling salesman problem, involving just seven cities, was solved. The importance was not solving that particular instance in itself but doing it differently; that particular instance would be pretty easy to solve on a computer as you have just 7! possibilities to check.

Re:Important nitpick

[Chris Burke]

Actually a particular instance of the traveling salesman problem, involving just seven cities, was solved.

Yes, but that's extremely impressive for a brand new method of computation. Most of the time, it's like "we've demonstrated a NAND operation so we're Turing complete!" or rarely "We have a full adder!" This was actual problem solving for a not-completely-trivial algorithm, which was pretty cool.

Of course it didn't "solve the traveling salesman problem" in a general sense -- that would be a feat well, well beyond demonstrating a new method of computation, it would revolutionize mathematics and algorithms. So yeah, I think it was obvious that "solved the traveling salesman problem for seven cities" meant "an instance of..." :)

The ultimate AI

[Anonymous Coward]

Scientists say they soon plan to inject sperm into a female human as means of using DNA to create a fully sentient and self-sufficient computer. What's next?!?!?

Computer virusses becoming trans-species

[JustNiz]

I understand virusses are not a living organism, but essentially just a chunk of DNA that happens to contain some element of self-replicating code, and that they effectively came into existence and propagate more or less by a mix of random mutation, natural selection and accident. The fact that this would work at all sounds unlikely but after millenia has in fact been successful enough that virusses are now everywhere in millions of differnet variants.

Consequently with the above research and others creating new and complex DNA structures that may become as ubiquitous as included in products in every home, aren't there serious risks around uncontrolled human exposure to this stuff?
I mean would it be possible for someone to become infected with some new man-made virus-like structure that actually originated/escaped from (say) their PC hardware? If so, is there anything that necessarily limits the potential for how bad the infection could be? I mean could this man-made DNA get accidentally integrated into humans own DNA and so modify the DNA of future generations? Could this technology ever potentially lead to a pandemic disease or even extinction scenario?