CBN Nano demonstrated mechanosynthesis on a silicon surface: molecular tools removed individual atoms, transferred carbon, and formed carbon-carbon bonds
CBN Nano demonstrated mechanosynthesis on a silicon surface: molecular tools removed individual atoms, transferred carbon, and formed carbon-carbon bonds
A series of arXiv papers describes inverted-mode STM, a version of the scanning tunneling microscope in which molecules on the substrate serve as both the “eye” and the chemical tool. The team worked with prepared silicon surfaces at cryogenic temperatures and in ultrahigh vacuum. The question is whether covalent reactions can be mechanically steered with atomic precision.
On 23 June, Death Is Nothing to Us published a breakdown of three CBN Nano papers. The topic looks narrow if you forget the old Drexler-Smalley dispute over molecular machines. Drexler argued that future tools would be able to bring reactive groups to specific atoms. Smalley replied with the images of “fat” and “sticky” fingers: chemistry is too complex, atoms cannot be grabbed like construction parts, and the reactive fragment would cling to the tool itself.
The new experiments test a narrower question: can a rigid molecule, placed at the right point, force a covalent reaction down a chosen path? In one experiment, the tool removed a hydrogen atom from a designated site in 27 of 28 attempts. In another, C₂ fragments, meaning pairs of carbon atoms, were transferred onto pre-prepared reactive sites and extended a chain through a new C-C bond. In a third, the team changed the structure of the tool, and it began pulling out silicon atoms: 63 of 63 times at 4 K and 100 of 100 times at 77 K.
The mechanosynthesis here works without any fantastical miniature arm. A chemical group is fixed inside a molecular framework, the microscope moves it with sub-angstrom precision, and the geometry of the approach makes one reaction outcome more likely than the others. After the operation, the surface is scanned again. In one case, the researchers saw an extra silicon atom in the working area and removed it in the next step. That is already a “check the result, find the defect, make a corrective move” scheme.
The limits are still severe. The experiments were run on clean Si(100), far outside biological tissue. They require vacuum, cold, a prepared surface, and slow sequential operations. The action set is tiny: remove hydrogen, transfer C₂, build a short carbon structure, remove silicon. There is still an enormous engineering distance from this to a molecular factory, medical nanorobots, or repair of a cryopatient.
For cryonics, the result lands in a narrow but important gap. Cryonics rests on a conditional wager: if the personally significant structure of the brain is preserved well enough, future technology may be able to repair damage after ischemia and cryopreservation. Brain vitrification without chemical fixation is already testing the first part of that wager: does the tissue ultrastructure remain after cooling? Mechanosynthesis bears on the second: does physics allow targeted molecular repair as a class of operation.
The answer is still small, but concrete. Positionally controlled covalent chemistry has moved from the books of Drexler, Freitas, and Merkle into a laboratory system: find an atomic site, bring in the tool, remove an atom, add a fragment, form a bond, verify the result, and sometimes correct an error.
The less ischemia, ice, toxicity, and destruction there is today, the less a future machine will have to reconstruct by guesswork. The best repair starts with less damage.
The next argument is already technical: can this vocabulary be expanded to thousands of reactions, rechargeable tools, automatic verification, parallel work, and an environment where this chemistry could have any bearing on a preserved brain?