Inverted‑mode STM achieves atom‑by‑atom carbon placement on hydrogen‑terminated silicon

What the authors claim

The preprint Atomically precise mechanosynthesis of carbon structures on hydrogenated Si(100) by inverted‑mode STM (arXiv:2605.27250) describes a procedure in which a scanning tunneling microscope (STM) operated in “inverted mode” extracts C₂ fragments from deposited organic molecules and deposits them onto deliberately created dangling‑bond sites on a H‑Si(100) surface. According to the authors, the technique enables:

1. Single‑site C₂ donation – a single carbon dimer placed at a chosen reactive site.
2. Patterned multi‑site donation – arrays of C₂ units written in a user‑defined geometry.
3. Stepwise polyyne growth – successive C–C bond formation that yields linear carbon chains up to several atoms long.

If reproducible, these capabilities would constitute a concrete implementation of the “mechanosynthesis” concept that has long been discussed in nanofabrication circles.

---

How the experiment works

Surface preparation

• A clean Si(100) wafer is hydrogen‑terminated, creating a surface where each Si atom is bonded to a hydrogen atom. This passivation renders the surface chemically inert.
• Using a conventional STM tip, the authors selectively desorb hydrogen atoms from chosen lattice sites, exposing Si dangling bonds that act as reactive anchors.

Inverted‑mode STM

Traditional STM measures tunneling current while maintaining a constant tip‑sample distance. In the inverted mode used here, the tip is deliberately pressed into the surface to a depth that allows mechanical interaction with adsorbed molecules. The authors cite earlier work on “mechanical manipulation” with STM (e.g., Science 2022, 376, 1234) as the conceptual basis.

Carbon source and transfer

• A thin film of a C₂‑rich precursor (the paper does not specify the exact molecule, but references to a “surface‑deposited organometallic complex”) is deposited on the H‑Si surface.
• When the tip contacts a precursor molecule, the mechanical force ruptures a specific bond, liberating a C₂ fragment.
• The fragment then migrates to the nearest dangling‑bond site, where it chemically bonds to the Si atom, forming a Si–C linkage.

Verification

The authors rely on a combination of low‑temperature STM imaging, scanning tunneling spectroscopy (STS), and ex‑situ Raman spectroscopy to confirm the presence of carbon dimers and longer polyynes. The reported images show bright protrusions at the programmed sites, and the Raman spectra display peaks near 2100 cm⁻¹ consistent with C≡C stretching.

---

What is actually new?

1. Mechanical extraction of a defined carbon fragment – Prior STM work has demonstrated atom manipulation (e.g., moving individual Si atoms on a Si surface) and tip‑induced bond breaking, but controlled release of a specific molecular fragment is rare. The paper provides the first experimental evidence that a C₂ unit can be detached and placed with sub‑nanometer precision.
2. Integration with a chemically inert substrate – Hydrogen‑terminated silicon is a standard platform for atomic‑scale lithography. Demonstrating that a carbon fragment can be grafted onto Si dangling bonds without destroying the surrounding H‑termination is a useful proof‑of‑concept for hybrid silicon‑carbon nanostructures.
3. Stepwise chain growth – By repeating the donation process on adjacent dangling bonds, the authors build short polyyne chains (up to four carbon atoms reported). This shows that the method can be iterated, a prerequisite for any programmable assembly line.

---

Limitations and open questions

Aspect	Observation	Implication
Yield	The paper reports successful C₂ placement in ~30 % of attempts for single‑site experiments. Multi‑site patterns have lower overall success rates.	The technique is still stochastic; scaling to larger arrays would require substantial improvement in reliability.
Throughput	Each C₂ transfer takes several seconds of tip‑positioning and force application.	Even a modest 10 × 10 pattern would require minutes of continuous STM operation, far from industrial relevance.
Chemical specificity	The precursor chemistry is not fully disclosed, and the authors note occasional side reactions that leave residual fragments on the surface.	Without a well‑characterized, clean source, reproducibility across labs will be difficult.
Structural verification	STM/STS provide indirect evidence; the Raman peaks could arise from adsorbed fragments rather than covalently bonded chains. No cross‑sectional transmission electron microscopy (TEM) or X‑ray photoelectron spectroscopy (XPS) data are presented.	Independent validation of the bonding motif is needed before the community accepts the claim of true Si–C covalent attachment.
Scalability to 3‑D	The method operates strictly on a planar Si(100) surface. Extending the approach to multilayer structures would require additional steps not addressed in the paper.	Current relevance is limited to 2‑D surface patterning; broader mechanosynthesis goals remain out of reach.

---

Context within the field

The idea of building devices atom by atom dates back to the 1990s, most famously with the IBM “atom‑by‑atom” writing of a xenon‑based logo on Ni(110) using STM. Since then, two main strands have emerged:

• Lithographic approaches such as hydrogen‑resist lithography on Si, which excel at defining dopant patterns but lack direct chemical control over the added species.
• Molecular‑scale manipulation, where individual atoms or small clusters are moved or chemically transformed with an STM tip.

The present work bridges these strands by coupling a lithographic pattern (hydrogen desorption) with a mechanical chemical reaction (C₂ donation). It therefore occupies a niche that has been largely unexplored.

---

Practical implications (if the method matures)

• Hybrid silicon‑carbon electronics – Covalently attached carbon chains could serve as molecular wires interfaced directly to silicon circuitry, potentially enabling ultra‑short interconnects.
• Quantum‑dot arrays – Precise placement of carbon clusters might be used to define localized electronic states for quantum information experiments.
• Catalyst design – Atomically defined Si–C motifs could act as model catalysts for surface reactions, offering a clean platform for mechanistic studies.

At present, all these applications remain speculative because the demonstrated process is far from the throughput and yield required for device fabrication.

---

Next steps for the research community

1. Publish detailed precursor chemistry – A reproducible synthesis route and full characterization of the C₂ donor molecule are essential.
2. Provide independent structural validation – High‑resolution TEM or XPS would settle the question of bond formation.
3. Improve reliability – Optimizing tip geometry, force control, and temperature could raise the single‑site success rate above 70 %.
4. Demonstrate functional devices – Even a simple two‑terminal test structure (e.g., a carbon‑linked Si nanowire) would move the work from proof‑of‑concept toward application.

---

Where to find the preprint

The full manuscript, supplementary information, and raw data are available on arXiv: https://arxiv.org/abs/2605.27250. Interested readers can also follow the authors’ updates on the project’s GitHub repository (link not provided in the paper).

---

This article is intended for readers familiar with nanoscale fabrication techniques and does not constitute endorsement of the reported results. The claims should be evaluated in the context of reproducibility and scalability challenges that have historically limited atom‑scale manufacturing.