Diamond-building Proposal

Radical nanotechnology has taken a big step closer to the lab.

One kind of molecular manufacturing uses mechanosynthesis: transferring a few atoms at a time from a mechanically positioned "tool tip" to a growing workpiece. The idea is to build programmable shapes by doing a few reactions many times in programmable positions. This way, just a few reactions can be used to build an almost unlimited range of shapes.

Robert Freitas, the nanomedicine expert, has been working on how to build diamond using mechanosynthesis. He has previously posted a proposal to find a comprehensive set of diamond-building reactions that could be used to build arbitrary shapes; he thinks this may need as few as six to ten tool tips. And he thinks these tips can be found in as little as five years and $5 million.

But that's theory, and now Freitas is talking practice. He's just posted a transcript of the talk he gave at the recent Foresight conference.

The talk covers a lot of ground in detail. First, he overviews his previously published work (with collaborators at Zyvex) simulating a two-atom transfer from a tool tip to a diamond surface ("dimer deposition"). Then, he goes through four steps that could be used to build a diamond mechanosynthesis tool. The idea is to synthesize a tool tip molecule, deposit it on a surface to orient it, add a large "handle," and end up with a tool tip molecule attached to a nanopositioner, having a reactive dimer that will bond to a diamond surface when positioned correctly.

Step 1: Synthesis of Capped Tooltip Molecule
Step 2: Attach Tooltip Molecule to Deposition Surface in Preferred Orientation
Step 3: Attach Handle Structure to Tooltip Molecule
Step 4: Separate Finished Tool from Deposition Surface

He doesn't say a lot about step 1, other than that they're working on it and similar molecules have been synthesized as well as found in nature. Sounds to me like they're waiting till they write a patent.

He presents at least two alternate pathways for steps 2 and 3. One way to attach a multi-micron "handle" to the molecule is simply to use a CVD diamond-growing process; this has already been experimentally demonstrated on other molecules. Very clever! Step 4 should be accomplished simply by pulling.

Once the tip is free of the surface and held by the manipulator, it can be used to test dimer-deposition reactions.

This all looks doable today. It's only a fraction of the way to diamondoid molecular manufacturing. But it's a very important fraction. If it works, it will demonstrate once and for all that mechanically guided diamond-building vacuum chemistry is feasible. The recipe is detailed enough that it's already hard to argue it can't work. The claim that diamond-building can't be used to build machine parts seems likely to lose most of its remaining credibility.

Chris

Comments

[jim moore]

There are several aspects to Robert’s proposal that I find quite striking.

- It provides a good outline on how to move from today’s technology to diamondiod nano-factories with out going through biological intermediate steps. (Any comments Dr. Jones?)
- The lab equipment needed to start this process is not that exotic. It looks like you need a Carbon Vapor Deposition Chamber, an AFM or STM, MEM’s based micromanipulator, electron microscope and a few other things. Now, a lot of the equipment may have to be modified for the particular task but the point is that are a number of people/ organizations that are experts with each type of equipment.
- During the fist couple of steps in the development phase you would need much less than one gram of the molecular tool handle. So you don’t need a large quantity of exotic chemical’s to do the work.
- Each step along the development pathway helps you build the next step.

It looks to me as if the “Freitas Process” is open to investigation and development by a very wide range of small actors (universities, moderate to large sized companies) not just the Major world powers.

[jim moore]

Also at that same nano-tech conference I had a very interesting lunch with Steve Sullivan who I believe works for a company called Nanoconductor. Steve told me about some work that they are doing. He believes that they have found a way to mechanically manipulate a sheet of graphite in such a way as to be able to "pinch" off a carbon nanotube. Essentially if this process works you should be able to make any type and any length of buckytube you want.

I think that it works something like this: If you fold the sheet of graphite in the right way you create two lines of strained carbon - carbon double bonds . You align one set of stained bonds with the other set of stained bonds and the system moves to a lower energy state. The result is you "pinch" off the buckytube from the sheet of graphite.

[Chris Phoenix, CRN]

It's not clear how that nanotube-building system would obtain the sheet of graphite, or align it. Such a task would be pretty far in advance of any technology I'm aware of. Also, I'm skeptical whether it can scale to making enough buckytubes for large-scale use like electrical conductors. A square meter of graphite is about 800 micrograms. A square meter per second is about 3 grams per hour. But I can't imagine how they could grow graphite by the square meter without grain boundaries.

Sheets of graphite (graphene) have recently been peeled off of graphite blocks. But these are small. If they can achieve alignment (possibly through some X-ray diffraction technique? if that even works on single sheets) then this might be useful for making buckytubes for computer chips and sensors. But not for large-scale production, unless I'm really missing something.

Chris

[jim moore]

Chris,
The system he described was rather simple. You take a very precisely built cylinder that has tiny and precise groves in it and carefully press and roll the cylinder over a graphite crystal. You pull the bucky tube off the cylinder and roll it up.

He said by this time next year they will have made and tested one of these cylinders. Time will tell.

To align sheets of graphite I would use something like pairs of STM's, X-rays are too energetic to use with out causing damage to your sample. Or maybe you could use and electrical field to align the crystal or sheet of graphite.

[Richard Jones]

It's good to see some concrete proposals, and particularly good to see schemes to extend the rather limited number of systems for positionally-specific mechanosynthesis that have been demonstrated so far. I'd suggest that you'd want to see whether you can achieve a deposition of the C2 group from the "tooltip" molecule using an STM before you embarked on the complicated business of trying to attach a "handle". It's stating the obvious to point out how important the "recharge" step is if you are ever going to make anything useful with the process.

[Chris Phoenix, CRN]

I think the point of the "handle" was to make it possible to pick up the tip with a MEMS gripper system. Another suggested way to pick up the small-molecule tooltip was to bond a reactive scanning probe tip to it. For example, I guess an AFM tip could be crashed/broken in vacuum, scanned in tapping mode to find the molecule, and pressed onto the molecule to attach the molecule to the AFM tip.

A possible third way, which Freitas didn't suggest, is to functionalize the tooltip molecule with thiol groups and pick it up with a gold STM tip.

I'm sure experimenters will do whichever is easiest and/or most likely to get them a steady stream of grant money. And I'm sure that as soon as they manage to manipulate an activated tool tip molecule by any means, they will try to do the dimer deposition.

Chris

[jim moore]

All graphite approach to MNT?

Sense we will need some form of graphite (graphene, Bucky tubes, bucky balls) in carbon based MNT any way, why not use single sheets of graphite (graphene) or carbon nano tubes as the starting raw material for an assembler / nano-factory?

-Fewer number of reactions. Instead of depositing every carbon atom modify a preexisting sheet or tube.

-If you combine precise mechanically induced strain in a sheet of graphene (or bucky tube) with positional placement of potentially reactive molecules ( or nano-stuctures) you may have enough “tools” to make a variety of nano-blocks. I was thinking something like: take sheets of graphene with a set size and send them through a stencil - stamp - stack -stick process to make the nano-blocks.
Stencil - adds a patterned monolayer.
Stamp - removes a pattern of carbon atoms from the sheet.
Stack - turns the two dimenstional pattern into a 3-D object.
Stick - joins the endges of the graphene sheets together.

-Graphite has a very different arrangement of electrons compared to diamond. In diamond all the eclectrons are locally bound this limits how diamond can interact with electro-magnetic fields. Graphititic material has what is known as a pi-bond system, each carbon atom has one electron that has been “de-localised” into a hexagonal “nano-circut”. Because the electrons are able to move about they can interact with a much wider variety of electro-magnetic fields. The exact structure of the graphititic material determins its EM interatctions, different stuctures have different EM interactions.

-Bucky balls have been put into a diamond anvil and have been crushed into diamond.

[Chris Phoenix, CRN]

"Modify a pre-existing sheet or tube". Hm... Sounds good at first, but the problem is, how do you make the precise shapes you need, and how do you grab and position them? Or how do you take a large random-shaped (or at least random-length) graphene object and cut it down to size, without leaving lots of messy bits?

Can buckyballs be used as building blocks? But they're not very close to graphene sheets, and cutting and joining them seems more complex than building from scratch.

If you know a reaction that can do the "Stamp" step--cut out only part of a graphene sheet according to adsorbed or bonded atoms/molecules--then that could help a lot.

Electron microscopes can be used to cut and weld buckytubes. I could see using that to build a first nano-fabricator. But I don't think it can be made to scale.

I'd be happy to be proved wrong on any of this.

Chris

[Martin Baldan]

This is great news! It seems that the bottleneck is to synthesize the DCB molecule. All the other steps look relatively straightforward. Could it be that molecular manufacturing is but one small molecule away?

IMHO, the order of steps to take after the first rudimentary tools are developped should be as follows:

1) Build a few dozen units of a standarised rechargeable handle.

2) Build a grid to store the handles, so that they can be picked, used and placed back in it, and then all of them recharged by placing the grid in the right environment (let's say acetylene).

3) Build a gripper which can be attached to an AFM tip so that it can easily pick, use and place the handles.

4) Build a simple MNT positioning device, so as to get rid of the AFM. No rod logic should be necessary for this.

5) Use this positioning device to build more positioning devices, exponentially.

6) Applications: small nanofactories to build larger ones, MNT computers and then all the rest.

Martin

[Tom Craver]

The race between the bottom-up intentional nanotech builders, and the random-assembly unintentional nanotech ground-breakers is shaping up nicely.

The former will try to create single-point assembly functionality, which they can then leverage to replicate into an array.

The latter are heading down a path of crude self-organizing patterned arrays with very simple molecular functionality at each point.. Their work should iteratively develop in complexity and precision until it can create simple assembler functionality at each point in the array.

Who will get to the nanofactory first? Maybe they'll meet halfway and cut off the last few years of development time for both.