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."

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: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: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:It's not "Cal Tech"

[cfa22]

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

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 [jyi.org])