Tabletop Fabrication

Scientists at the Georgia Institute of Technology are announcing a three-dimensional microfabrication technique that uses a unique class of light-activated molecules to selectively initiate chemical reactions within polymers and other materials, which could provide an efficient way to produce complex structures with sub-micron features.

"We have developed a disruptive platform technology that we believe will provide broad new capabilities," said Seth Marder, a professor in Georgia Tech's School of Chemistry and Biochemistry. "We believe this technique provides a real competitive advantage for making complicated three-dimensional microstructures."

Producing each structure now requires about 25 seconds, but increases in speed could make mass-production feasible.

"We are working to integrate the technologies and develop a system that should be able to operate at a thousand times the throughput of the current system," said Georgia Tech's Joseph Perry. "A single 3D fabrication system should be able to generate about a million individual device structures per day. With a production facility using a number of fabrication systems, there is potential for certain types of mass production."

The researchers envision tabletop fabrication machines that would use a computer-generated design system to laser write the desired structures. See the full announcement here.

Comments

[Brett Bellmore]

I notice that they're claiming horizontal resolution to a third of a wavelength. That's rather amazing... Though not quite impossible, given the nonlinear nature of double photon reactions. Very impressive!

[Mr. Farlops]

So I wonder what this thing will do for MEMS once it becomes commonplace. Can you imagine generating a million tiny, electopolymer (Excuse me while I make up some fictious wonder-plastic.)robots waiting for control chips to be implanted? A zillion MEMS built to order.

[Brett Bellmore]

Nothing particularly fictictious about that; Polymer's generally do exibit at least some mechanical response to electrical fields, and if it's not a big response, well, you don't WANT a manipulator for molecular manufacturing to move very far.

I imagine a chip with hundreds of thousands of Stewart platforms manufactured onto its surface, and particular reagents bonded to different platforms. Simply place the chip face down on a working surface, and use some of the platforms to walk it around, in order to bring the specific reagent sites you want to use within reach of the point you're working at. Not a terrifically fast or efficient way to do molecular level manufacturing, but it could do as a bootstrap technique.

[Mr. Farlops]

Yes, going with your stewart platform idea, imagine using a refined version of this fabricator to build millions of nanowalkers each armed with a scanning probe to jog atoms around on a surface. Perhaps with such a system and a lot of patience we could start building blocks of stuff layer by layer. A sort of cumbersome proof of concept behind mechanosysthesis.

But we are getting ahead of ourselves.

1. The fabricator has to show that it can build such MEMS.
2. We have to figure out a way to miniaturize scanning probe rigs down to the size of these MEMS
3. We have to refine our scanning probe techniques to work under less extreme conditions (Something reliable at tempuratures higher than liquid nitrogen for example. And with a much wider variety of atoms, molecules and bonding sites.).
4. We have to work out the software to control all these little robots.

[Brett Bellmore]

Liquid Nitrogen is no real concern. It's one of the cheapest fluids around, easy to obtain, easy to store. A household nanofactory which had to be as cold as THAT would be perfectly practical, though there would be some product design constraints, to avoid stress fractures due to differential thermal expansion to room temperature.

Creating a MEMS STM system is, in my opinion, no show stopper. The real challenge is learning to create and repeatably place on the tip a suitable set of reactant molecules, AND to understand their behavior well enough to design a sequence of reactions to construct a more practical assembler. The first assembler is going to be a slow, unreliable hack job, not a practical system for anything except bootstrapping to a better system.