Mechanosynthesis: From Theory to Experiment

Inverted‑Mode STM and the Path to Atomically Precise Fabrication

For decades, molecular nanotechnology has occupied an ambiguous position in discussions about cryonics. The ability to manipulate matter with atomic precision would seem to provide the kind of technology required to repair severe ischemic and cryopreservation injuries. Yet detailed proposals for molecular repair have remained largely theoretical and, following the Drexler–Smalley debate, have often been dismissed as chemically implausible.

The central issue in that debate was not whether molecular manufacturing would require many reactions, specialized tools, error correction, automation, and systems integration. Eric Drexler never denied this. The dispute concerned a more fundamental question: do physics and chemistry permit reliable mechanical and positional control of molecular reactions, particularly outside the aqueous, enzyme-mediated environment of biology?

A recent series of papers from CBN Nano Technologies begins to answer that question experimentally. Researchers have developed surface-bound molecular tools, introduced inverted-mode scanning tunneling microscopy (inverted-mode STM), removed selected hydrogen and silicon atoms, deposited carbon fragments at chosen sites, and formed carbon-carbon bonds through repeated mechanical operations.

Robert Freitas and Ralph Merkle are coauthors and conceptual contributors to the three principal inverted-mode STM papers. Their involvement connects the experimental program to more than two decades of theoretical work on mechanosynthetic tools, carbon-dimer placement, reaction trajectories, failure modes, and minimum reaction sets for atomically precise manufacturing.

The experiments do not constitute a molecular assembler or demonstrate medical nanorobots. Their significance is narrower but still considerable: designed molecular tools can now perform a small but growing set of reproducible covalent operations at selected atomic sites.

From Moving Atoms to Controlled Chemistry

Scanning tunneling microscopes have been used to move atoms and molecules since the pioneering IBM experiments of the late 1980s. Subsequent work enabled atomic patterning, hydrogen removal from silicon, dopant placement, and surface reactions. The deeper challenge was not whether atoms could be moved, but whether covalent chemistry could become predictable and reproducible enough to support manufacturing.

In conventional STM, the apex atoms of a sharp metallic probe participate in both imaging and manipulation. Their exact structure and chemistry may be uncertain, and the tip can change during an experiment. This becomes a serious limitation when the probe is intended to function as a standardized chemical tool.

Inverted-mode STM reverses the arrangement. The moving probe presents a comparatively broad crystalline silicon surface. Tall, rigid molecules attached to the opposing surface act as both imaging probes and chemical reagents. As one of these molecules scans beneath the silicon surface, the electronic interaction between the tip and the surface maps the topography of the build site, creating a sharp atomic-scale image.

This makes it possible to inspect a location, perform a reaction, examine the result, and revisit the same site with another tool. The innovation is therefore not simply improved imaging. It is an experimental platform for repeated, inspected chemical operations at the atomic scale.

The precursor paper, “Molecular Tools for Non-Planar Surface Chemistry,” introduced the molecular architecture behind this work. The tools attach to a surface through several legs while presenting a protected reactive group away from it. A rigid central structure constrains their orientation, and activation exposes the chemically active end.

Not every deposited molecule adopts the required configuration, so suitable tools must still be identified by their imaging signatures. Nevertheless, the central principle was established: a molecule can function as a localized chemical instrument.

This is important because atomically precise manufacturing would not depend on one universal nanoscale hand. Instead, much like a modern multi-tool CNC machine or a robotic assembly line, it relies on a modular repertoire of specialized tool-tips. It requires distinct tools for adding material, removing it, transferring particular groups, forming selected bonds, and verifying the results.

Three Experimental Breakthroughs

Mechanical Hydrogen Abstraction

The first dedicated inverted-mode STM paper demonstrated the removal of individual hydrogen atoms from a hydrogen-passivated silicon surface.

A molecular tool was activated by removing a protective iodine atom and aligned with a selected hydrogen atom on the silicon surface. During the transfer itself, the electrical bias was reduced to zero and the two surfaces were brought together mechanically. At sufficiently close range, the hydrogen transferred to the molecular tool, leaving a reactive dangling bond at the chosen position.

The researchers reported successful hydrogen abstraction in 27 of 28 trials.

Hydrogen removal from silicon was not unprecedented. What was new was the combination of a designed molecular reagent, a selected atomic target, a mechanically initiated reaction, and inspection of both the altered tool and the modified surface.

Adding Carbon and Forming New Bonds

The second paper moved from subtraction to additive fabrication.

Pairs of reactive sites were prepared on hydrogenated silicon. Activated tools then transferred two-carbon fragments to those selected locations, producing the intended initial structure in more than 90 percent of attempts.

Multiple carbon units could be placed in predetermined patterns. More significantly, a second carbon fragment could be added to one already present, producing a new carbon-carbon bond and a four-atom carbon structure.

The longest continuous carbon structure demonstrated was only four atoms. This is not a diamond lattice or a new bulk carbon material. But it established a sequence central to positional mechanosynthesis: deposit a fragment, retain it at a chosen site, return with another fragment, and form a larger covalently bonded product.

The yield declined as the sequence became more complicated. A reaction that succeeds 90 percent of the time is impressive chemistry but inadequate for structures requiring thousands or millions of operations unless failures can be detected, corrected, or avoided. This in situ imaging capability offers the beginnings of such a feedback process because each product can be inspected before the next operation.

A Specialized Tool for Removing Silicon

The third paper explored both carbon donation and silicon-atom abstraction on an unpassivated silicon surface.

With the original tool, either reaction could occur: the carbon fragment might be deposited, or a silicon atom might be pulled from the surface. The researchers then redesigned the molecule by changing its central atom and shortening its supporting legs.

The resulting tool strongly favored silicon abstraction. Under the reported conditions, it removed silicon in all 63 tested operations at 4 K and all 100 tested operations at 77 K, with no carbon donations observed.

This result applies to one tool, one broad target geometry, and a limited set of trajectories. Its deeper significance is that changing the molecular architecture changed the preferred chemical outcome. Selectivity could be engineered into the tool itself.

The researchers used it to remove individual silicon atoms and create small vacancy patterns. When one operation left an unwanted silicon atom on the surface, they identified and removed it in a subsequent step. This is a rudimentary form of error correction: inspect the result, identify the defect, perform a corrective operation, and reinspect the site.

Operation at 77 K (-196°C) is also notable. It remains cryogenic and requires ultra-high vacuum, but it is considerably less demanding than operation near 4 K (-269°C).

Expanding the Reaction Set

Conference abstracts indicate that the next stage of the program concerns hydrogen-containing carbon groups such as C₂H and C₂H₂.

This represents more than another carbon-deposition experiment. A bare two-carbon fragment is symmetric, whereas an asymmetric group like C2​H possesses inherent chemical directionality. Controlling not just the placement coordinate, but the structural orientation of these asymmetric building blocks, is a prerequisite for constructing complex, non-planar molecular components.

Other manuscripts in preparation concern the characterization of carbon radicals and alternative methods for producing the same surface structures. The research program therefore extends beyond the manipulation experiments themselves to molecular synthesis, surface chemistry, theoretical modeling, product identification, and tool design.

From Theory to Experiment: Drexler, Freitas, and Merkle

The current research has several related intellectual origins.

In Nanosystems: Molecular Machinery, Manufacturing, and Computation (1992) Drexler described how rigid molecular machinery could position reactive groups, constrain their motion, and favor selected reaction pathways without requiring miniature hands that grasp isolated atoms.

The relationship with Freitas and Merkle is more direct. Both are coauthors of the three inverted-mode STM papers and are credited with contributing to their conceptual development. The present work is partly the experimental continuation of a research program in which they remain active participants.

Beginning in the early 2000s, Freitas, Merkle, and their collaborators studied proposed tools for transferring carbon dimers to selected sites on diamond surfaces. Their theoretical and computational work examined tool structures, reaction pathways, hydrogen abstraction, tool recharging, failure modes, and the minimum operations needed to construct increasingly complex carbon structures. They subsequently advanced this program through the Nanofactory Collaboration and patents associated with CBN Nano Technologies.

Revisiting the Drexler–Smalley debate

The Drexler–Smalley debate concerned whether chemistry permits the positional control required for molecular manufacturing. Richard Smalley argued that mechanically directed synthesis encountered “fat fingers” and “sticky fingers”: too many atoms would need to be controlled simultaneously, and a reactive fragment held by a tool could not reliably be released at its destination. Drexler replied that his proposals involved positioning bonded reactive groups, not manipulating each atom with a separate mechanical finger. Geometry, constrained motion, and competing reaction pathways would determine the result.

The inverted-mode STM experiments provide substantial support for Drexler on this central point. A molecular tool can be positioned over a selected site and used to alter a particular covalent structure. Carbon can be held during positioning and then transferred to silicon. Other tools can remove hydrogen or silicon atoms. Redesigning the tool can change the dominant reaction from mixed donation and abstraction to highly selective abstraction. In these limited systems, the “fat finger” and “sticky finger” problems have proved to be design problems rather than fundamental prohibitions.

This does not establish that a general molecular manufacturing system can now be built. The reaction repertoire remains small, the operations are serial, and the tools are altered during use. But these limitations do not show that mechanically directed positional chemistry is incoherent. They define an engineering program involving a larger reaction set, tool reloading, improved reliability, automation, error correction, parallel operation, and the construction of useful three-dimensional products. Drexler’s core claim—that physics and chemistry permit molecularly precise mechanical control—has now received significant experimental support.

Mechanosynthesis and Cryonics Revival

The cryogenic, tightly controlled conditions of inverted-mode STM may seem to limit its relevance to manufacturing or medicine. The experiments use prepared silicon surfaces, ultra-high vacuum, and liquid-helium or liquid-nitrogen temperatures.

But the conditions needed to develop or manufacture a machine (or its parts) are not necessarily those in which the finished machine must operate. Semiconductor fabrication relies on vacuum chambers, clean rooms, elevated temperatures, and carefully controlled chemistry, while the resulting devices function in ordinary environments.

For cryonics, low-temperature operation may also be less alien than it first appears. In Cryostasis Revival, Freitas proposes that the initial scanning and repair of a cryonics patient would deliberately occur at cryogenic temperatures to avoid disturbing an already fragile molecular state.

Low temperatures suppress diffusion, spontaneous reactions, autolysis, and other structural changes. A patient would be examined and stabilized without first exposing the surviving structure to normothermic biological temperatures. Premature warming could reactivate damaging processes before the tissue had been mapped or repaired.

Cryogenic operation may therefore be useful, or even necessary, during the earliest stages of revival. The first objective would not be to restore normal metabolism immediately, but to stabilize and record the information embodied in the preserved structure before controlled warming.

The present experiments establish several categories of operation relevant to molecular reconstruction: identifying atomic sites, removing selected atoms, adding fragments, forming covalent bonds, inspecting products, and correcting errors.

Future machines would not necessarily need to reproduce the exact environment or architecture of the current STM apparatus. They might be constructed under highly controlled conditions but designed to function in solid cryogenic tissue, aqueous environments, or transitional states during gradual warming.

The distance between manipulating a few atoms on silicon and repairing a human brain remains enormous. A revival system would also require sensing, computation, transport, energy management, waste removal, and the preservation or reliable inference of identity-critical neural information, topics extensively discussed in Cryostasis Revival.

One component of the broader reconstruction scenario—positionally controlled covalent chemistry using designed molecular tools—now has a stronger experimental foundation.

This is not a reason for complacency about current cryopreservation injury. The possibility of future molecular repair does not justify avoidable ischemia, ice formation, cryoprotectant toxicity, dehydration, or fracturing. Better preservation remains the most reliable way to reduce the burden on future repair technologies.

An Emerging Experimental Science

This research invites two opposite exaggerations. One is that molecular assemblers and medical nanorobots are now inevitable. The other is that the experiments are irrelevant because they are slow, cryogenic, and limited to a few atoms.

The more useful question is whether the experimental vocabulary continues to expand. Can tools be reloaded? Can more elements and functional groups be transferred? Can longer, branched, ring-shaped, or three-dimensional structures be built? Can errors be corrected automatically? Can the process be parallelized and independently replicated? Can it produce assemblers and machines intended for environments different from those required for their construction?

The answers will determine whether inverted-mode STM remains a specialized form of surface science or develops into a more general platform for atomically precise fabrication.

Researchers can now identify an atomic site, position a molecular tool, remove an atom, deposit a molecular fragment, form a bond, inspect the outcome, and attempt a correction. This is not yet molecular manufacturing, but it is credible evidence that an experimental science of positional mechanosynthesis is beginning to emerge.

References

Barrera et al. “Inverted-Mode Scanning Tunneling Microscopy for Atomically Precise Fabrication.” 2025.

Blue et al. “Towards Atom-by-Atom Fabrication: Mechanosynthetic Donation and Abstraction.” 2026.

Cowie et al. “Atomically Precise Mechanosynthesis of Carbon Structures on Hydrogenated Si(100) by Inverted-Mode STM.” 2026.

Drexler, K. Eric. Nanosystems: Molecular Machinery, Manufacturing, and Computation. 1992.

Drexler, K. Eric, and Richard E. Smalley. “Nanotechnology: Drexler and Smalley Make the Case For and Against ‘Molecular Assemblers.’” 2003.

Freitas, Robert A., Jr. Cryostasis Revival: The Recovery of Cryonics Patients Through Nanomedicine. 2022.

Freitas and Merkle. “A Minimal Toolset for Positional Diamond Mechanosynthesis.” 2008.

Huff et al., “Molecular Tools for Non‑Planar Surface Chemistry,” 2025.

McCallum et al. “(Hetero)adamantane Synthesis: A Triple Alkylation Reaction.” 2026.