In 1931, Hückel introduced the electron counting protocol to account for the peculiar thermodynamic stability of the family of cyclic and planar hydrocarbons featuring delocalized [4n + 2] π-electron systems—now called Hückel aromaticity1. Later, Breslow proposed that cyclic molecules with 4n π electrons are antiaromatic2. Notably, the relationship between the [4n + 2]/4n-electron and its aromatic/antiaromatic nature can invert depending on the spin-state. That is, conjugated 4n-electron molecules in the lowest triplet state (T1) or the singlet excited state (S1) exhibit aromatic properties—Baird aromaticity3.

Nowadays, the concept of the aromaticity/antiaromaticity based on the [4n + 2]/4n-electron counting rule as well as the spin-state is not confined to organic compounds but is widely applied to various molecular systems ranging from organometallics to inorganic species. Moreover, in addition to archetypal π aromatics, other types such as σ-, δ-, and φ-aromatics are proposed, which has expanded the implication of the concept of aromaticity4. Irrespective of the system, the classical [4n + 2]/4n-electron counting approach still remains the standard method for the initial instant evaluation of the aromatic nature of molecules. In this context, are all planar compounds with 4n π systems in the singlet ground state always supposed to be anti-aromatic? In other words, can aromatic singlet molecules with a globally 4n π system exist?

Is it possible to split a 4n π system into sub [4n + 2] π systems?

In 2010, Wang, Boldyrev, and colleagues reported the B19 cluster, which was generated by a laser vaporization method and detected by photoelectron spectroscopy (PES)5. Computational studies revealed that the global minimum (C2v) of B19 exhibits a planar geometry with one central boron atom surrounded by a pentagonal B5 ring in the first coordination sphere and a B13 ring in the second coordination sphere (Fig. 1a). B19 possesses in total 12 π electrons, suggesting that it is antiaromatic according to the [4n + 2]/4n-electron counting criteria. Nevertheless, its π aromatic nature is confirmed by the nucleus-independent chemical shift (NICSzz) value of −14.9 ppm at 0.6 Å above the B19 plane as well as molecular orbital analysis. This can be rationalized by considering it as a concentric, doubly π aromatic system, in which two π electrons are delocalized over the central pentagonal B5 unit, and ten π electrons are delocalized in a circle between the central B6 fragment and the peripheral boron atoms, respectively. Consequently, each π-electron circulation simultaneously obeys the [4n + 2] Hückel rule—an example of concentric aromaticity. Interestingly, Heine, Merino, and colleagues revealed the fluxional behaviour of B19. That is, the internal and external boron rings rotate nearly freely with respect to each other, like a Wankel motor6.

Fig. 1: Aromatic boron species with globally 4n π electrons.
figure 1

a Molecular structures and schematic drawings of concentric doubly π-aromatic systems. b Molecular structures and schematic drawings of ribbon aromatic systems. c Synthesis of isolable B4C2 carborane (Dip = 2,6-diisopropylphenyl. Me = methyl).

In 2011, the same group generated a B18 species by the method described above and analysed it with PES7. According to ab initio calculations, one of the global minimum structures of B18 (C3v), which is only slightly more stable by 1.6 kcal mol−1 than the other isomer (Cs), is quasi-planar and consists of an inner triangular B6 unit and a peripheral B12 ring (Fig. 1a). B18 involves the 6c–2e π bond delocalized over the inner B6 moiety and five π bonds (three 5c–2e and two 18c–2e) delocalized between the inner six B atoms and the peripheral ring. Accordingly, analogous to B19, B18 is also concentric doubly π aromatic. These studies illustrate that boron clusters even with globally 4n π electrons in the singlet state may in fact be aromatic by splitting them into several cyclically delocalized [4n + 2] π-electron systems. A subsequently emerging question is whether this way of dividing π electrons is limited to concentric circles.

Can we cut 4n π systems into small strips?

In 2012, the group of Wang and colleagues reported a series of dehydrogenated boron clusters H2Bx (x = 7–12) that were characterized by PES and computational studies8. The global minima of the clusters H2Bx have ladder-like double-chain structures capped by a hydrogen atom at each terminal boron. The chemical bonding analysis revealed that the π bonding patterns in this class of the clusters bear an analogy to polyenes. Among them, the clusters H2B7, H2B8, and H2B9 are found to possess a 4π-electron system, indicative of antiaromatic character. Shortly thereafter, Li and colleagues reported that the B4 rhombus in the planar clusters of H2B4 (D2h), H2B8 (C2h), and H2B12 (C2h), is not only equivalent to a C=C double bond unit in the corresponding CnHn+2 (n = 2, 4, 6) hydrocarbons but possesses aromatic character (Fig. 1b)9. For instance, each terminal B4 rhombus in H2B8 is covered by a delocalized 4c–2e π bond and the NICS(0)πzz is –22.9 ppm. Thus, breaking the 4π system into two 2π strips may equip the cluster with an aromatic feature—called ribbon aromaticity.

The aforementioned systems (B19, B19, H2B8) are transient species generated in situ and characterized only by PES. Is the chemical synthesis of aromatic 4n π electron systems possible?

How to isolate and structurally characterize such species

Recently, our group developed a carborane B4C2R4 by the reaction of tetraatomic boron zero species B4(0) with isonitriles (Fig. 1c)10,11. B4C2R4 was obtained as orange crystals in 23% isolated yields. The solid-state molecular structure determined by an X-ray diffractometry shows that the skeletal B4C2 framework is perfectly planar with a double-chain structure, similar to H2Bx−/0. Computational investigations show that the B4C2 core of B4C2R4 involves 4π electrons, as well as, the NICS value at 1.0 Å above the CBB three-membered ring is −11.8 ppm, in line with the NICS value of −15.0 ppm of the parent derivative B4C2H4. This indicates that the B4C2 core bears two delocalized 3c–2e π bonds (Fig. 1b), leading to the ribbon aromatics. In B4C2R4, (H2B8 as well) the central B4 rhombus has a 4c–2e σ-bond, forming the consecutive π- and σ-delocalized system8,10. In 2015, the group of Lu and Li proposed that such a unique conjugation system is expected to appear in longer double-chains as well12. Our work demonstrates that the synthesis of bottleable ribbon aromatic species is feasible by employing the proper substituents and synthetic approach.

Outlook

Mathematically, the union of the numbers represented by [4n + 2] gives 4m (m = 2n + 1). Accordingly, as far as the construction of [4n + 2] π sub-units is plausible, even compounds with 4n π could have a chance to be aromatic, indicating that an instant judgement of aromaticity of molecules only by the Hückel rule needs to be avoided. The open questions remaining are as follows: Apart from concentric and ribbon aromatics highlighted above, what would other systems exist? What about other p-block elements-containing systems? Could we establish the protocol to predict the local aromaticity of 4n π molecules from their structure?

Both theoretical and experimental elaboration will lead to the answers to those questions. Computationally, further studies to explore molecules bearing the globally 4n and locally delocalized [4n + 2]-electron systems can be anticipated. Synthetically, isolation of such species remains highly challenging but we believe that it is attainable to develop them by the proper approaches. For instance, by employing the judiciously selected Lewis bases as the supporting ligands, an isolable version of the concentric aromatic species and more can be prepared. Not only are these species fundamentally significant synthetic targets, but their peculiar electronic structures should contribute to the development of nanomaterials.