Atomically precise mechanosynthesis of carbon structures on hydrogenated Silicon

5/29/2026 at 7:50:45 AM

This is the experimental science behind the Eric Drexler mechanosynthesis->nanomachines ideas from the 90's and 2000's. We got decades of pop sci-fi and memes (and the Borg) out of this idea, and a hazy vision of the future singularity where AI would control the material world.

Well, we didn't need AI to start down the path. The core 'can this be done' science is done. Yes, we can control chemical reactions by manipulating reactant position.

Next the question is how do we scale it? We need LOTS of manipulators. We're... not actually great at building lots of manipulators. We might use this atomically precise tech to build manipulators and have them make copies of themselves.

Or we might leverage other technologies like protein engineering, DNA origami, and nanoelectronics to make less effective manipulators that we could build today and use to bootstrap toward the nanotech that CAN build everything atom-by-atom.

Either way, things are accelerating. Finally.

by jeremybarton

5/27/2026 at 5:09:55 PM

This seems to be the galena cat whisker stage of developing Drexler's "eutactic chemistry". It may be a parlor trick forever, or it might be the humble beginnings of mechanosynthesis, the first working tool-tips for a Drexler arm.

https://wiki.reprap.org/wiki/Drexler_Arm

by MarkusQ

5/27/2026 at 4:05:34 PM

At the current time I see this is front page and has reasonable upvotes, but not one comment.

My guess is that this is highly useful for either:

* Building chips (but it's on a silicon surface, not with silicon)

* Carbon nanotubes (looking forward to a space elevator, folks, one day ;))

* It's a more notable movement forward in nano construction in general and I just know too little to understand its impact.

Given the zero comments to date: would someone who knows this area like to accept an invitation to share something really interesting about this paper, please? :)

by vintagedave

5/27/2026 at 4:34:10 PM

Inside a vacuum chamber cooled at very low temperatures, they move a scanning tunneling microscopy probe over the surface of a silicon wafer, and with it, in the locations that they choose, they can build some carbon atom structures attached to the silicon surface, atom by atom, in a reproducible way.

This kind of control of atomic structures, where the position of every individual atom can be controlled, is very impressive, but for now this is pure research. Any practical applications are more than a decade in the future.

Techniques derived from this might be used some day either for making some kind of integrated circuits or for making some kind of masks for the production of such circuits, but there is no way to use this directly for the production of some material needed in great quantities, like nanotubes for a space elevator.

For a material needed in great quantities, if you build it atom by atom, there must be billions of such building processes done concurrently, like in a living being, in order to achieve an acceptable growth rate. For now, they use a microscopy probe to attach the atoms one by one and you could not make a sufficient number of these that work concurrently.

by adrian_b

5/27/2026 at 5:49:40 PM

This is an experimental demonstration of the technology that makes Neal Stephenson's "Diamond Age" possible. They place an individual carbon atom pair(dimer) on silicon. Silicon has the same crystal structure as diamond, so adding carbon in this manner is similar to building diamond. There is potential for eventually making 3D atomically precise structures of diamond with this tech although that is far off.

This is a very, very basic demonstration, but it's showing for the first time that atoms can be controllably and programmably covalently bonded in something that can probably be extended to 3D. There was a lot of skepticism that this process was possible at all, so an experimental demonstration is important

by gene-h

5/27/2026 at 7:27:26 PM

I think it would be useful for experimental construction of individual new molecules, for example for providing data for molecular simulation codes.

by pfdietz

5/28/2026 at 9:11:58 AM

precision is to aprox 1/3 the diameter of a carbon² molecule, and are effectivly printing matter at the atomic level, but too slowly for manufacturing at any scale large enough for a practicle use other than perhaps, creating test circuits for.....more testing and ,maybe labeling and logos

by metalman

5/27/2026 at 6:50:51 PM

On my cork board I'll paste this up and run a red string to https://arxiv.org/abs/2605.05914

and put a post-it between them with COMPUTONIUM???? written on it

Can we please keep it together for a modest number of decades longer please.

by aaroninsf

5/27/2026 at 5:12:02 PM

Off-topic, these days information just goes in circles from subreddits, X, YouTube, and Hacker News to countless secondary sources, and then back again to the same original sources.

by num42