The new compounds were not the product of a bespoke stabilization strategy. The chemists set out to probe the reactivity of a reactive diboracumulene, [(CAAC)BB(CAAC)], in which two boron atoms are joined by a multiple bond and each carries a cyclic (alkyl)(amino)carbene ligand. Such B–B multiple-bonded compounds are known to cleave the carbon–oxygen and carbon–sulfur bonds of carbonyl and thiocarbonyl compounds, including carbon dioxide and tautomeric thioketones, while with acetone they instead trigger carbon–hydrogen activation and cyclization. Expecting yet another C=O cleavage, the team treated the diboracumulene at room temperature with excess benzophenone, 4,4′-dimethoxybenzophenone and the amide N,N-dimethylformamide. Instead, each reaction delivered an oxoborane in greater than 50% isolated yield—three compounds, 1a–c, each bearing borylenyl groups of the form –B(CR2)2.

All three oxoboranes were isolated as orange or yellow crystalline solids and are stable at room temperature, a property that sets them apart from every previously reported free oxoborane. Their defining feature is an exceptionally short and strong boron–oxygen bond. Single-crystal X-ray diffraction gave B–O distances of 1.208(4) Å for 1a, 1.212(2) Å for 1b and 1.217(4) Å for 1c—essentially the same as the 1.210(3) Å bond of the platinum oxoboryl complex trans-[Pt(BO)Br(PCy3)2], probably the shortest known B–O bond until now, and slightly shorter than the 1.232(4) Å bond of the Lewis-acid-stabilized oxoborane Mes*BO(AlCl3). The B≡O units are essentially linear, with angles of 176.2(2), 174.0(2) and 179.5(1) degrees respectively.

Spectroscopy reinforced the picture of unusually robust triple bonds. The solid-state infrared spectra of 1a–c show the diagnostic pair of B–O stretching bands for the boron-10 and boron-11 isotopomers: 1,922 and 1,860 cm−1 for 1a, 1,921 and 1,860 cm−1 for 1b, and 1,886 and 1,827 cm−1 for 1c. These values sit well above those of Lewis-base-stabilized oxoboranes (around 1,600 cm−1), above the platinum oxoboryl complexes (1,853/1,797 and 1,849/1,793 cm−1 measured in dichloromethane) and at or above the Lewis-acid-stabilized compound III (1,830 cm−1), while remaining slightly below the 1,975 cm−1 recorded for the transient oxoborane PhB≡O frozen at 10 K. The authors conclude that the B–O bonds in 1a and 1b are probably the strongest ever observed in stable, isolated compounds.

Computational work at the ωB97XD/Def2-TZVP level of theory corroborated the experimental data. Calculated B–O bond lengths of 1.211, 1.211 and 1.215 Å match the crystallographic values closely, and calculated NMR shifts for the two boron nuclei in each molecule agree well with the measured 11B signals (for example, 1a: observed 24.7 and 16.0 ppm versus computed 21.2 and 13.2 ppm). Natural resonance theory bond indices for the B≡O bonds—2.914, 2.905 and 2.535 for 1a–c—confirm near-triple-bond character, while the calculations indicate that the triple bond is considerably ionic in nature, with oxygen contributing more than boron to the bonding orbitals. By contrast, the B–B bonds, at 1.669(5), 1.660(2) and 1.659(5) Å in the solid state, are close to typical single bonds and carry negligible π character, so the molecules are best described as discrete, non-delocalized oxoboryl and borylenyl units connected by a single B–B bond.

Why these molecules survive when their predecessors did not comes down to electronics rather than steric armor. The environment around the oxoboryl unit in 1a–c is not unusually shielded, yet the compounds are stable. The key, the authors argue, is the electron-rich borylenyl substituent –B(CR2)(CAAC), which acts as a strong σ donor. This σ basicity reduces the positive charge at the boron atom of the B≡O unit (computed as +0.97 for 1a and 1b and +0.92 for 1c, compared with +1.09 for transient PhBO and +1.12 for Mes*BO), making it less susceptible to nucleophilic attack. A related effect operates in the platinum oxoboryl complex II, whose electron-rich [–Pt(SPh)(PCy3)2] fragment lowers the corresponding charge to +0.67. Attempts to quantify the Lewis acidity at the oxoborane boron atom by the Gutmann-Beckett and Müller methods failed, probably because of the steric shielding there.

Two further computational signatures distinguish the stable compounds from transient species such as PhBO and Mes*BO. First, the lowest unoccupied molecular orbital with perceptible π*BO character lies far higher in energy for 1a–c and II (+3.13, +3.28, +2.88 and +3.19 eV) than for the transient oxoboranes (+0.20 and +0.52 eV), making these orbitals reluctant to accept electrons from incoming reagents. Second, maps of the electrostatic potential on the molecular van der Waals surfaces show that in PhBO and Mes*BO the negative potential is concentrated at the oxygen atom with a single local minimum, marking a highly polarized, reactive B≡O unit. In 1a–c and II the negative potential is dispersed across multiple minima over the whole BBO framework, tempering bond polarization and B≡O-focused reactivity.

The synthesis was not entirely free of surprises. During the preparation of 1a, a few blue crystals of a minor product appeared, which X-ray diffraction identified as a cyclic diradical, compound 2. Running the reaction at 80 °C instead of room temperature flipped the product ratio, delivering 66% of the diradical and 33% of the oxoborane. No interconversion between the two could be observed on further heating, and adding one equivalent of CAAC or other Lewis bases to 1a did not generate the diradical, indicating that 1a and 2 form through distinct pathways. High-level DLPNO-NEVPT2/CASSCF multireference calculations showed that 2 has an open-shell singlet ground state with substantial biradical character (y0 = 0.65), lying 1.55 kcal mol−1 below the triplet state. In 2 the C=O bond of benzophenone has been cleaved and B–O bonds formed, but the B–B bond has also been cleaved, which probably prevents oxoborane formation. No analogous diradical appeared in the reactions that produced 1b and 1c.

Despite their stability, the new oxoboranes are not inert. Initial reactivity tests with amines revealed B–O-centred chemistry. Treating a solution of 1a with two molar equivalents of pyrrolidine brought about rapid 1,2-hydroamination of the B≡O triple bond, yielding the (amino)(hydroxy)boryl-substituted boraalkene 3 in 87% yield; X-ray diffraction showed a B–B single bond of 1.721 Å. Substituting 2-aminopropane for pyrrolidine diverted the reaction to a spirocyclic product, 4, structurally similar to the diradical 2 but with the CAAC unit bridging the two boron atoms. A computed mechanism for the formation of 1a traces a plausible route: precoordination of the benzophenone oxygen to one boron atom, a [2 + 2] cycloaddition across the B=B bond, carbon–oxygen bond cleavage, and finally ejection of the CAAC unit bound to the oxygen-bearing boron, with the overall reaction computed to be strongly exergonic (ΔG° = −34.1 kcal mol−1).

The achievement opens a field that has been closed for more than sixty years. The reported route, which amounts to simply combining a diboracumulene with ordinary carbonyl-bearing reagents such as benzophenone or dimethylformamide under argon at room temperature, removes that bottleneck. Beyond the oxoboranes themselves, the work highlights electron-rich borylenyl groups as powerful σ-donating substituents that can stabilize other hypocoordinate main-group fragments, a strategy the authors suggest could be applied more broadly. With stable, isolable B≡O compounds now in hand, the full reactivity landscape of these long-sought species can finally be mapped.

Subject of Research: Synthesis and isolation of the first stable, base- and acid-free oxoboranes with strong B≡O triple bonds

Article Title: Synthesis and isolation of stable oxoboranes

Article References: Lubczyk, L., Markl, C., Jain, M., Kar, S., Kupfer, T., Dewhurst, R. D., Phukan, A. K., & Braunschweig, H. (2026). Synthesis and isolation of stable oxoboranes. Nature Chemistry. https://doi.org/10.1038/s41557-026-02272-w

Image Credits: AI Generated

DOI: 10.1038/s41557-026-02272-w

Keywords: oxoboranes, boron chemistry, main-group chemistry, B≡O triple bond, diboracumulene, borylene, X-ray crystallography, density functional theory, diradical, hydroamination, Nature Chemistry, reactive intermediates

A team of chemists at Seoul National University led by Seung Youn Hong has now demonstrated a conceptually different route to the same goal, one that does not require bending a sulfur atom at all. Writing in Nature Catalysis, Woojin Lee, Jason Kim and colleagues describe a spirocyclic organosulfur scaffold in which four sulfur atoms are anchored to a single carbon centre, a geometric arrangement that forces the sulfur p orbitals into direct spatial overlap with one another. The result is a highest-occupied molecular orbital that is no longer localized on a single atom but spread across four sulfur centres simultaneously, and whose energy is substantially elevated compared with ordinary sulfides. In effect, the molecule manufactures its own reactivity not by distorting a single atom but by making several atoms share a common electronic frontier.

The underlying phenomenon, known as through-space orbital interaction, has been familiar to physical organic chemists for decades under the name spiroconjugation, and it has been exploited in optoelectronic materials, where spiro-linked frameworks tune the electronic properties of organic semiconductors. What the Seoul team has done is to transport this concept from materials science into catalysis. In their spirocyclic design, the rigid carbon framework holds the four sulfur atoms in a fixed three-dimensional relationship, ensuring that their lone-pair orbitals approach each other closely enough to mix. Computational analyses and sulfur K-edge X-ray absorption near-edge structure measurements, performed in collaboration with Jongwoo Lim’s group, confirmed that the highest-occupied molecular orbital is a genuine four-centred orbital with markedly raised energy, a signature that prototypical acyclic sulfides do not display.

Elevated orbital energy translates directly into chemical power. The highest-occupied molecular orbital governs a molecule’s nucleophilicity, its willingness to donate electron density into an incoming electrophile. A four-atom orbital spread across sulfur centres means that the sulfide form of the catalyst is an unusually potent nucleophile, far more electron-rich than a conventional thioether of comparable structure. The researchers showed that this enhanced nucleophilicity allows the spirocyclic sulfide to attack N-halosuccinimides, common halogen-transfer reagents, with pronounced ease, forming a sulfonium intermediate that carries an activated halogen ready for delivery to an arene substrate.

What makes the system a true catalyst rather than a stoichiometric reagent is the reversibility of this chemistry. Once the sulfonium species transfers its halogen to an aromatic ring, the catalyst is restored to its sulfide state and can re-enter the cycle. Mechanistic studies, including kinetic analyses designed to establish reaction orders, revealed a striking form of multi-atom cooperativity: the same four-centred orbital architecture that makes the sulfide state hyper-nucleophilic also endows the corresponding sulfonium state with an unusual capacity for halogen transfer. Both halves of the catalytic cycle, reduction and oxidation in a formal sense, are amplified by the shared orbital, which is precisely the kind of dual behaviour that geometrically constrained phosphorus catalysts achieve through distortion of a single centre.

The flagship application demonstrated by the team is electrophilic aromatic halogenation, one of the oldest and most industrially important reactions in organic chemistry. Aryl chlorides and bromides are ubiquitous in pharmaceuticals and agrochemicals; the chlorine atom alone appears in a large fraction of blockbuster drugs, where it modulates potency, metabolic stability and membrane permeability. Yet installing halogens onto complex, electron-poor or sterically congested arenes remains difficult, and existing organocatalytic halogenation methods each carry limitations in substrate scope or reactivity. The spirocyclic organosulfur catalyst proved capable of halogenating arenes that are otherwise hard to functionalize, exhibiting a level of activity that clearly distinguishes it from prototypical organosulfur compounds such as simple sulfides and even from previously reported specialized halogenation catalysts.

Of particular interest to the pharmaceutical community is the catalyst’s performance in late-stage functionalization. The researchers subjected drug-like molecules, scaffolds decorated with multiple sensitive functional groups, to the halogenation conditions and found that the spirocyclic catalyst could install halogen atoms selectively onto aromatic positions without disturbing the rest of the molecule. This kind of chemoselective, late-stage diversification is a prized capability in medicinal chemistry, where a single advanced intermediate may serve as the branching point for dozens of analogues in a structure-activity relationship study. A metal-free, organocatalytic platform that tolerates complex molecular architecture expands the synthetic toolkit available for such campaigns and avoids the toxicity and environmental concerns associated with many metal-based halogenation catalysts.

The study also carries conceptual weight for the broader field of main-group chemistry. Over the past fifteen years, chemists such as those working on geometrically constrained phosphorus(III) compounds and low-valent bismuth platforms have shown that main-group elements can mimic transition metals in catalysis, performing oxidative addition, redox cycling and bond activation once thought to be the exclusive province of the d-block. That programme, however, has been built almost entirely on the geometric-distortion principle, which is inherently element-specific. By showing that through-space orbital interactions can achieve an analogous electronic effect in a divalent chalcogen, the Seoul team has added a second, orthogonal design axis. The strategy is in principle transferable to other divalent elements and to other spirocyclic architectures, suggesting a generalizable route to multicentred main-group catalysts that does not depend on forcing unusual bond angles.

The experimental evidence underpinning the claims is unusually thorough for a conceptual advance. Crystallographic data for three representative spirocyclic compounds were deposited with the Cambridge Crystallographic Data Centre, establishing the solid-state structures and the spatial relationships between the sulfur atoms. Sulfur K-edge XANES spectroscopy, supported by simulations, probed the electronic structure of the sulfur centres in both the sulfide and sulfonium states, while quantum-chemical calculations mapped the composition and energy of the four-centred highest-occupied molecular orbital. Energy decomposition analyses traced the halogen-transfer step to the cooperative involvement of multiple sulfur atoms, providing a mechanistic rationale for why the multicentred catalyst outperforms its single-centred counterparts.

For a reaction as old as aromatic halogenation, the arrival of a genuinely new catalytic principle is noteworthy in itself. But the deeper significance of the work may lie in the demonstration that molecular orbitals can be engineered through architecture rather than distortion, that the spatial arrangement of identical atoms around a shared framework is itself a designable variable in catalyst development. If the spirocyclic strategy generalizes as the authors suggest, chemists may soon be building multicentred main-group catalysts for a wide range of transformations, from redox chemistry to bond activation, by simply deciding which atoms to bring within orbital reach of one another. The periodic table, it appears, still has unexploited modes of cooperation waiting to be switched on.

Subject of Research: Through-space orbital interactions in a spirocyclic tetrathiide scaffold that generate a four-centred high-energy HOMO for multicentred organosulfur catalysis

Article Title: Harnessing through-space orbital interactions for multicentred organosulfur catalysis

Article References: Lee, W., Kim, J., Lee, H., Hwang, Y., Park, J., Lee, S., Lim, J., & Hong, S. Y. (2026). Harnessing through-space orbital interactions for multicentred organosulfur catalysis. Nature Catalysis. https://doi.org/10.1038/s41929-026-01602-y

Image Credits: AI Generated

DOI: 10.1038/s41929-026-01602-y

Keywords: organosulfur catalysis, through-space orbital interactions, spiroconjugation, frontier molecular orbitals, main-group chemistry, electrophilic aromatic halogenation, multi-atom cooperativity, late-stage functionalization, sulfonium intermediates, organocatalysis, divalent chalcogens, Nature Catalysis