A plethora of beautiful new aromatic compounds, with increasing intricacy and complexity, are reported every year. On the other hand, when one looks back at the seminal work in the field, one sometimes discovers surprisingly simple systems that remain unexplored and mysterious, even for decades. This happened to us when considering the pioneering papers of Franz Sondheimer and co-workers, describing chemistry of annulenes.1,2 [n]Annulenes are fully pi-conjugated polyenes, where the number n indicates ring size, for example: [18]annulene is (CH)18. The gargantuan efforts of Sondheimer’s team in the 1960s verified the validity of Hückel’s rule: annulenes are aromatic if they contain 4n+2 electrons in the circuit, and anti-aromatic when there are 4n electrons (where n is a positive integer).
[18]Annulene is one of the iconic molecules of organic chemistry, yet surprisingly underexplored. Perhaps the myth of its instability and the nontrivial synthesis are what hindered researchers from further exploration. In 1973, Oth, Woo and Sondheimer reported that reduction of [18]annulene with potassium metal leads to an anti-aromatic dianion, with enormous differences between the magnetic shielding inside and outside of the ring (ca. 30 ppm by 1H NMR) reflecting its strong anti-aromatic character.3 Based only on the symmetry and integration of the poorly resolved spectrum, they concluded that [18]annulene dianion exists in two structural forms (Fig. 1). While the first geometry seems reasonable, the second one looks strange, potentially with substantial steric clashes between the inner protons, so we decided to re-investigate this archetypal compound.
Figure 1. Geometry of previously postulated species involved in reduction of [18]annulene with potassium metal together with reported 1H NMR chemical shift values.3
The easiest way to get unequivocal structural information would be to grow single crystals and determine the geometry from X-ray diffraction. That’s why my PhD supervisor, Harry, contacted Marina Petrukhina — her group in Albany (NY, US) has outstanding expertise in crystallization and coordination chemistry of highly air-sensitive anions.
I synthesised [18]annulene, travelled with it to Albany, and got paired up with a PhD student from Marina’s laboratory, Yikun Zhu. Our collaboration and mutual understanding made the research efficient and genuinely exciting. We worked together on the alkali metal reduction — at first, very cautiously, because according to Sondheimer, the anion was expected to decompose above 0 °C. Another difficulty was stopping the reaction at an appropriate stage — the original work did not even mention the beautiful colour changes accompanying the reduction, so we were not sure what to expect, and when to stop the reaction: is it monoanion? Dianion? Or did it already decompose? Fortunately, all the stars aligned for us, and we managed to get suitable crystals (using lithium metal as a reducing agent) in only two weeks (Fig. 2).
Figure 2. Flame-sealed crystallization ampules (left) and crystals of tetra-anion under microscope (right).
Fortunately, Marina’s group was awarded extra synchrotron beamtime at Argonne National Laboratory, so analysis of our precious material became possible even on very small crystals. Zheng Wei, the departmental crystallographer, took our sample to Chicago for a night shift. I remember being woken up by a few notifications on my phone and reading the exciting news: X-ray diffraction analysis indicated not only that instead of dianion we crystallised an unexpected tetraanion, but also that it has a totally different geometry to that assigned 50 years ago and, on the top of that, it forms a lithium-intercalated sandwich (Fig. 3,4)! Up to this point, the only hydrocarbon capable of forming similar sandwiched complexes was corannulene.4 This was the first of many Eureka! moments during this project.
Figure 3. Geometry of the tetra-anion and crystal structure of the self-assembled sandwiched complex. Color code: blue - Li, grey - C, red - O. Hydrogen atoms were removed for clarity.

Figure 4. 3D-printed model of the annulene sandwich, together with lithium metal in oil and yellow solution of [18]annulene, in front of an X-ray diffractometer.
Nuclear magnetic resonance (NMR) experiments in solution indicated that lithium metal yields an anti-aromatic di-anion and aromatic tetra-anion; with potassium, the reaction does not reach the tetra-anionic stage. Although we did not manage to force the di-anion to crystallize, the symmetry of the NMR signals clearly indicated that its shape is identical to that of the tetra-anion. Contrary to the original paper, we found that these anions are stable at room temperature, provided they are protected from oxygen and moisture.
After two months in Albany, I came back to Oxford, synthesised more material and performed a series of variable temperature NMR experiments. They showed that the dianion is flexible and undergoes dynamic conformational changes in solution, indicated by well-resolved multiplets recorded only at or below –70 °C. In contrast,the tetra-anion is rigid, and its NMR spectra do not exhibit significant temperature-dependence, remaining sharp even at 55 °C.
The last (but not least) eureka moment came at the final stage of the project. I still had a few milligrams of material left and realized that due to the resemblance between [18]annulene and corannulene tetra-anions (both having five internal bays coordinating lithium cations in their sandwich complexes) it would be interesting to mix them together and see, whether there is any trace of a heteroleptic sandwich. Looking at the NMR spectrum made me absolutely stunned — it suggested selective formation of the mixed compound! I ran directly into Harry’s office with printed spectra, and we discussed the results. In less than a week, crystals were formed, and I was able to measure them and solve the structure, getting tangible evidence for formation of a mixed sandwich (Fig. 5).
Figure 5. Geometry of corannulene and [18]annulene tetra-anions and fragment of a crystal structure of the heteroleptic sandwich.
The reasons for the drastic change in geometry during the reduction were unclear until our postdoc, Igor Rončević, performed a series of state-of-the-art theoretical calculations. According to the picture drawn from the relative energies and electronic structure of the different possible geometries, the observed change of shape minimizes the repulsion of the negative charge accumulated on the apical positions of the ion.
Our research demonstrates that the [18]annulene dianion exists as a single isomer, in contrast with previous reports. The reduction process results in an unusual change of geometry when going from the neutral molecule to the anti-aromatic dianion and (previously unknown) aromatic tetraanion. Self-assembly of the tetra-anion with lithium cations leads to a metallocene-like, diannulene sandwich complex and a heteroleptic (mixed) sandwich with a corannulene tetra-anion. This research re-writes history and shows potential of using higher annulenes as multicoordinating ligands.
References:
1. Sondheimer, F. The annulenes. P. Roy. Soc. A - Math. Phy. 297, 173–204 (1967).
2. Sondheimer, F., Wolovsky, R. & Amiel, Y. Unsaturated macrocyclic compounds. XXIII. The synthesis of the fully conjugated macrocyclic polyenes cycloöctadecanonaene ([18]annulene), cyclotetracosadodecaene ([24]annulene), and cyclotriacontapentadecaene ([30]annulene). J. Am. Chem. Soc. 84, 274–284 (1962).
3. Oth, J. F. M., Woo, E. P. & Sondheimer, F. Unsaturated macrocyclic compounds. LXXXIX. Dianion of [18]annulene. J. Am. Chem. Soc. 95, 7337–7345 (1973).
4. Zabula, A. V., Filatov, A. S., Spisak, S. N., Rogachev, A. Yu. & Petrukhina, M. A. Main group metal sandwich: Five lithium cations jammed between two corannulene tetraanion decks. Science 333, 1008–1011 (2011).
Follow the Topic
-
Nature Chemistry
A monthly journal dedicated to publishing high-quality papers that describe the most significant and cutting-edge research in all areas of chemistry, reflecting the traditional core subjects of analytical, inorganic, organic and physical chemistry.
Please sign in or register for FREE
If you are a registered user on Research Communities by Springer Nature, please sign in