Above: Examples of subtractive and additive mechanosynthesis under IM-STM/UHV conditions. HAT between activated 19 (chemisorbed on a Si(100)-2×1 surface) and H:Si(100)-2×1 SPC apex (A) and possible reaction outcomes between activated 19 (chemisorbed on a Si(100)-2×1 surface) and Si(100)-2×1 SPC apex including additive (C2-transfer) and subtractive (Si-atom abstraction) mechanosynthesis (B). BDEs calculated using ωB97X-D/Def2-TZVPP of model compounds 1-ethynyl-1-germaadamantane (model of molecular tool) and ethynyl-tris(trimethylsilyl)silane (model of surface-bound acetylene). See the pre-print for more.
Available as a preprint: https://chemrxiv.org/doi/full/10.26434/chemrxiv.15002728/v2
I described it as "Prelude to Foundational" on an X post, as the entire process that resulted in "Atomically precise mechanosynthesis of carbon structures on hydrogenated Si(100) by inverted-mode STM" (APM-IM-STM) was long, involved, and exploratory — the ground-floor search for the most viable conditions for performing the reported work by our SPM experts, theoreticians, and synthetic chemists at CBN Nano Technologies (CBNNT). To that end, a considerable improvement in adamantane cage chemistry — including the generation of functionalized (hetero)adamantane cages containing carbon and germanium bridgehead atoms (with the option for more!) — was realized by the amazing CBNNT synthetic chemistry team in collaboration with Brilliant Matters.
My small contribution (to v2) provided additional context for the applications in our current hardware setup – the one that provided the data for both inverted-mode STM (IM-STM) and additive mechanosynthesis. Folks with a history tracking the technical studies in the MNT/AMM/APM community likely have seen papers exploring molecular tool designs for, specifically, carbon dimer transfers. This simple C₂ feedstock was the basis for proposing fabrication of diamond and related materials based on mechanosynthetic principles, from basic tool designs all the way through what became 1:40–2:40 of the famed Nanofactory Animation.
Whereas playing with the electronic structure of strained π-bonds to promote dimer transfer was key to the predicted behavior of DC10c (I now host a local copy of the PDF, since the journal version is no longer retrievable), most of the molecular transfer tool papers employed germanium bridgehead atoms for this task. The Ge–C bond is a fair bit weaker than the C–C bond, making dimer transfer highly favorable in configurations where the in-process build is ready to accept those dimers into an extended carbon lattice through basic radical coupling. While tin and lead would appear to be reasonable candidates based on their placement in the periodic table (C–Si–Ge–Sn–Pb), they were computationally demanding (and certainly were back in the early 2000s to get right), were prone to much more interesting chemistry owing to their large, polarizable electron shells (which risks the chemistry becoming more involved, including possible intramolecular reactions on delivery), and were certainly not the favorite elements to do organic chemistry with whenever the conversation came up with our chemists or with potential collaborators (which, as a chemical grandchild of John Verkade, I get). For many of the details on the use of germanium, see the bibliography and available papers at molecularassembler.com/Nanofactory/AnnBibDMS.htm. And, from this paper:
"Herein lies the significance of the functionalized germanium bridgehead as a first synthetic target for exploring additive mechanosynthesis through molecular transfer: bridgehead germaadamantanes provide a position for covalent bonding of species that can be activated for chemical transformations with far greater chemical tractability and stability than a comparable tin or lead analogue."
Fast-forward nearly 20 years, and the Ge bridgehead has survived a number of assessments of candidate tools to remain the delivery atom of choice, from Rob and Ralph's A Minimal Toolset for Positional Diamond Mechanosynthesis all the way to synthesis and first demonstrations. From all that work, this paper reports the significant improvements in yield and significant reduction in reaction steps for a convergent, triple-alkylation route to (hetero)adamantanes with interchangeable bridgeheads under mild conditions, turning two decades of Ge-bridgehead tool design into quantities sufficient for our mechanosynthetic exploration.
Things did get interesting on the long path to the first demonstrations. As is apparent from the APM-IM-STM paper, we are clearly jogging before running (from a purely diamondoid perspective, anyway) by working with a more tractable material in the form of Si(100) instead of diamond or any other carbon surface.
And here is the issue awaiting sufficient address. Bond-dissociation enthalpies for Ge–C and Si–C become uncomfortably close across an ethynyl (our C₂ dimer) fragment. As a result, every factor affecting bond strength at either position — dopant identity and placement on either side of the junction, current, defects, angle, and the residual theoretical uncertainty in the bonds themselves — becomes a variable driving the scission outcome. That, in turn, hints that additional levels of control are needed if the bridgehead, feedstock, and workpiece are to behave as they do in ongoing work and the desired outcomes (perhaps more philosophical than chemical — demonstrating a capability more than a specific product) are to be achieved.
Upcoming work (already discussed in presentations at conferences; I won't disclose details here prior to preprints being available, but am pleased to report such disclosures are imminent) explores just how uncomfortably close the bonds relevant to mechanosynthesis are in the current material/molecular-tool combination. Lessons learned from that process to follow.
Authors: Terry McCallum, Sam Rohe, Mathieu Morin, Hsin Yao Su, Samuel W. J. Shields, Kashif Tanveer, Marius Mamone, Nicolas Zindy, Damian G. Allis, Aru Joy Hill, and Michael Drew
ABSTRACT: (Hetero)adamantanes containing tetrel bridgeheads are important structures in surface and chemical sciences. Syntheses of diamondoid structures generally occur in a linear fashion, often using harsh conditions. These limitations make it difficult to access diverse (hetero)adamantane-based molecules with interchangeable functional groups for broad applications. The discovery of an expedient and convergent methodology for the synthesis of diamondoid scaffolds is described. The convenient synthetic route proceeding via a triple alkylation reaction represents a breakthrough in accessing the adamantane molecular class. Applications of this transformation to diamondoid tripods containing tetrel bridgeheads (C- and Ge-) were explored.
