A tiny change in the vacuum environment—atomic hydrogen from a hot filament—can stop a two-dimensional framework from growing. But the same chemical interference can also become a precise way to make monomers, dimers, and molecular hexamers from one precursor. Targeted synthesis of materials with designed properties or electronic structure is a central goal of materials research.

For ordered covalent organic frameworks, limiting chemical synthesis to two dimensions is especially challenging. On-surface synthesis offers one route: rigid molecular precursors are vapor-deposited onto a generally metallic substrate, which facilitates bond formation between molecular units.

This method allows covalent organic frameworks to be designed with specific lattice structures, functional groups, and electronic structure. Here, the focus is heterotriangulene two-dimensional polymers from tribromo-substituted dimethylmethylene-bridged triphenylamine monomer precursors.

On-surface synthesis is valuable for creating low-dimensional covalent organic frameworks, but it has limitations because effectively irreversible covalent bonds make framework extent and order highly dependent on deposition parameters and environmental factors. The presence of atomic hydrogen during deposition has been shown to inhibit carbon–carbon bond formation, so time-of-flight secondary-ion mass spectroscopy is used to provide conclusive evidence for hydrogen’s inhibitory effect on framework growth.

The study also examines whether atomic hydrogen can control surface chemistry, enabling single- and multimonomer compounds from a single precursor source, and follows their electronic structures with scanning tunneling microscopy and spectroscopy. Figure one maps the on-surface reaction of tribromo-DTPA on heated Ag(111).

At five hundred three to five hundred seventy-three kelvin, debromination produces surface-stabilized DTPA radicals; in high vacuum, they undergo Ullmann-type coupling to form a DTPA covalent organic framework, while an atomic-hydrogen environment can terminate some sites with carbon–hydrogen bonds, producing monomers, dimers, and other SAM fragments.

The figure matters because it visually connects deposition conditions with whether polymer growth continues or is interrupted. The precursor molecules were deposited from heated crystallized monomers onto temperature-controlled silver one-one-one substrates. On the surface, methyl groups point perpendicular to the surface and dominate scanning tunneling microscopy images, forming triangular molecular shapes.

The experiments were conducted in a turbo-pumped sample-introduction chamber, or load-lock, with an ambient pressure of two to three times ten to the minus eight millibar in high vacuum. These conditions are cautionary, yet they promise an additional level of control over on-surface synthesis.

Figure 2 compares STM topographs of DTPA deposited on Ag(111) under hydrogen-backfilled and high-vacuum conditions, while varying whether a remote cracking filament is off or on. The images show distinct surface products, including honeycomb-like molecular arrangements and regions labeled SAM, with the patterns changing across filament treatments and temperatures.

These observations matter because they support the authors’ proposal that atomic hydrogen from the filament inhibits on-surface carbon–carbon coupling and COF formation, while pre-exposure can leave the surface acting as a hydrogen reservoir. The top row of Figure 2 supports the hypothesis that atomic hydrogen produced at the remote cracking filament inhibits on-surface carbon–carbon bond formation and therefore covalent organic framework formation.

The source of atomic hydrogen was confirmed as thermal cracking of hydrogen gas at the hot filament, rather than energetic electrons or ions produced by accelerating potentials in the ion gauge. Figure three compares positive-ion TOF-SIMS spectra for DTPA and OTPA monolayers on heated gold surfaces.

In panel A, DTPA produces groups of peaks spaced by roughly fifteen mass units, consistent with fragmentation involving its bridging methyl groups during secondary-ion generation. Panel B instead shows a prominent OTPA peak at two hundred eighty-seven point zero six atomic mass units, with smaller isotope-related side peaks, supporting the assignment of the monomer ion and the absence of methyl substituents.

The mass spectrum shows vanishingly small intensities at masses corresponding to the tribromo precursor or any other partially brominated DTPA species. Below the monomer-DTPA mass, fragmentation from cleavage of increasing numbers of bridging methyl groups produces evenly spaced peak sets.

Within the exact-mass peak set, the largest peak at three hundred sixty-five point five atomic mass units corresponds to fully hydrogenated DTPA positive ion. Despite metastable-state complications at lower masses, the mass spectrum indicates that the prepared monomer self-assembled monolayers contain hydrogen-terminated DTPA molecules: hydrogen has replaced bromine in the precursor molecules.

The presence of atomic hydrogen inhibits the Ullmann reaction between organic molecules. By controlling deposition parameters, that inhibition can be used to synthesize a useful sequence of oligomers from the same tribromo DTPA precursors.

Figure 4 shows the variety of compounds and their self-assembled monolayers formed by deposition in controlled reducing environments. The topographs emphasize characteristic molecular structures and their assembly, rather than the entire surface distribution, which depends on deposition parameters and is statistical in nature.

Figure four compares scanning tunneling microscopy topographs with structural models for four DTPA assemblies: the covalent organic framework, hexamer, dimer or trimer, and monomer structures. The top row shows experimental images, while the enlarged views below overlay models to connect the observed patterns with molecular arrangements.

This matters because it presents the COF as one endpoint of a progression of oligomer structures formed under controlled reducing conditions, allowing the authors to relate assembly structure to synthesis conditions. As filament temperature, cracking rate, or atomic-hydrogen flux increases, the fraction of monomers increases while the percentage of molecules in the hexamer self-assembly decreases.

The reducing environment’s effectiveness can be quantified as the number of newly bound hydrogen atoms per DTPA, deduced as three minus the number of aryl–aryl bonds per monomer, as shown in Table 1. Table one reports how oligomer species are distributed on the silver one-one-one surface at three cracking-filament emission currents.

As emission increases from zero point zero four to one point zero milliampere, the table records changes in one-mer, two-mer, and six-mer percentages, alongside the inferred hydrogen-per-DTPA value, from one point seven to two point five. The authors use these statistics to show how atomic hydrogen during deposition can tune oligomer formation from the same precursor.

Figure five follows the electronic evolution from monomer and dimer, through a hexamer macrocycle, to the two-dimensional COF. Experimental STS spectra in panels A and C are compared with freestanding DFT density-of-states calculations in panel B, showing occupied levels in blue and unoccupied levels in red.

The occupied peaks progressively split and broaden into COF bands, while the unoccupied states are already closely spaced in the monomer and evolve more complexly; the labeled alpha one, alpha two, and beta peaks track features with similar spatial distributions. Figure 5A and Figure 5C display experimental filled-state and unfilled-state scanning tunneling spectra for monomer, dimer, hexamer macrocycle, and covalent organic framework islands.

Figure 5B shows density of states calculated at the density-functional-theory PBE level for the freestanding oligomers and framework. For the highest occupied molecular orbital levels, experiment and theory show peak splittings consistent with a linear combination of molecular orbitals, with one molecular orbital contributed by each monomer.

The tight-binding model reproduces the basic scanning tunneling spectroscopy and density-of-states results, including the hexamer degeneracies and slight asymmetry in the hexamer and framework peak energies caused by next-nearest-neighbor interactions. Peak beta may be influenced by the metal surface state, but its spatial distribution implies that it derives mainly from the adsorbed molecular layer.

Identifying beta as the lowest unoccupied molecular orbital peak allows a more accurate bandgap to be extracted. For the covalent organic framework experiments, the derived bandgap is one point eight five plus or minus zero point ten electron volts, nearly identical to density-functional-theory results for the freestanding framework.

That close agreement may be fortuitous, because the calculated bandgap is expected to decrease somewhat when a silver one-one-one substrate is included beneath the framework. Atomic hydrogen has a disruptive effect during on-surface synthesis of covalent organic frameworks, and the hydrogenated products quantitatively confirmed here impede Ullmann coupling.

Even a remote cracking filament and modest hydrogen pressure can create a sufficiently reducing environment to dramatically inhibit carbon–carbon bonding in ambient high vacuum, and the same effect could limit the size and quality of two-dimensional framework crystals in ultrahigh vacuum.

For large-area two-dimensional frameworks in vacuum, background hydrogen must be minimized and bare filaments should be cold during framework synthesis. The same inhibitory processes can be controlled and exploited to produce monomers, dimers, hexamers, and, to a lesser extent, three-, four-, and five-mers from a single monomer precursor, organized into self-assembled monolayers.

Figure six links structure to spectroscopy across a dimer SAM, hexamer SAM, and DTPA COF. Each row pairs an STM topograph with a color-coded map, where colors identify regions whose spectra resemble the characteristic K-means cluster spectra plotted at right.

The recurring alpha one, beta, and alpha two features can therefore be compared spatially, helping the authors distinguish molecule-dominant signals from substrate-influenced electronic states. Atomic hydrogen is both a hidden obstacle and a controllable reagent: it terminates reactive sites, suppresses COF growth, and lets one precursor produce a sequence of oligomers whose electronic structures evolve toward the two-dimensional COF.