Chemistry matters. Join us to get the news you need.

If you have an ACS member number, please enter it here so we can link this account to your membership. (optional)

ACS values your privacy. By submitting your information, you are gaining access to C&EN and subscribing to our weekly newsletter. We use the information you provide to make your reading experience better, and we will never sell your data to third party members.


Physical Chemistry

Aromaticity For All

Chemists argue over how the sacred concept of aromaticity should be invoked

by Stephen K. Ritter
February 23, 2015 | APPEARED IN VOLUME 93, ISSUE 8

Credit: Alexander Boldyrev/Lai-Sheng Wang
The calculated chemical bonding patterns of benzene and the B9 cluster indicate that both molecules are aromatic , with B9- both σ and π aromatic; ON is the electron bond occupation number.

Read Boldyrev and Wang’s response to Hoffmann’s article, a rebuttal by Hoffmann, and perspectives from other researchers, then add your thoughts at

Aromaticity is one of chemistry’s oldest concepts to describe the behavior of molecules. Chemists created it 150 years ago to help visualize and explain the bonding, structure, and reactivity of benzene.

Many chemists today hold aromaticity sacred and defend using the concept only for benzene, its many derivatives, and related heterocyclic compounds. Others believe the concept is so fundamental to chemistry that it should also be used to help explain the bonding, structure, and reactivity in any type of molecule, including fleeting inorganic compounds.

More on this story

Aromaticity For All

This repurposing of aromaticity is, to some traditionalists, the equivalent of claim jumping and is asking for a fight. In fact, theoretical chemist and Nobel Laureate Roald Hoffmann of Cornell University threw down a gauntlet last month when he published a historical perspective on the discovery and understanding of aromaticity in his regular column in the magazine American Scientist (2015, DOI: 10.1511/2015.112.18). In the article, Hoffmann lamented that some chemists had lately been stepping out of bounds by using aromaticity to describe bonding in planar and three-dimensional inorganic compounds.

Hoffmann referred to these chemists as “purveyors of hype.” Those purveyors, namely Alexander I. Boldyrev of Utah State University and Lai-Sheng Wang of Brown University, took Hoffmann’s comments as a terse indictment of their work and picked up his gauntlet, defending their usage of aromaticity. Their written response, a rebuttal by Hoffmann, and commentaries from other researchers in the field are available in full on C&EN Online (

Historically, chemists have viewed the characteristic stability of aromatic compounds as stemming from delocalized electrons in π bonds about the ring. These bonds form from molecular orbitals that overlap at two points, above and below the plane of the ring, and they must contain by definition 4n + 2 π electrons. But Boldyrev, Wang, and others suggest that compounds can also exhibit aromaticity involving other bond types, such as σ and δ bonds, which have a single point and four points of orbital overlap, respectively. They also suggest there can be mixtures of different types of aromaticity in the same molecule.

Hoffmann writes that “an inflation of hype” threatens the traditional, “beautiful” concept of aromaticity. “A century and a half after the remarkable suggestion of the cyclic structure of benzene, the conceptual value of aromaticity—so useful, so chemical—is in a way dissolving in that hype.”

Hoffmann says that many of the compounds produced by Boldyrev, Wang, and others “have precious little chance” of being made on a significant scale. As part of his argument, he puts forward two stringent criteria for aromatic molecules: They should not be overly reactive, and they should be bench-stable or bottleable.

Boldyrev and Wang have been chipping away at that confining notion over the past 15 years. Boldyrev is a computational chemist whose group determines whether imagined molecules are physically plausible, and Wang is an experimentalist whose group attempts to make them. The researchers generate a molecular beam of these species by vaporizing a disk containing the elements of interest with a laser, selecting out the target molecules via mass spectrometry, and then confirming their existence in the gas phase with photoelectron spectroscopy.

In 2001, Boldyrev and Wang reported the square-planar all-metal aromatic cluster Al42–. They contend that Al42– fulfills all the criteria for aromaticity, including 4n + 2 (n = 0) π electrons, equivalent Al–Al bonds, and high resonance stabilization energy. “If it walks like a duck, quacks like a duck, and looks like a duck, it must be a duck!” Boldyrev and Wang write in their response. “We do not see any ‘hype’ in the characterization of Al42– as being aromatic.”

Boldyrev and Wang subsequently made a series of planar boron clusters that they compare with polyaromatic hydrocarbons. For example, they argue that the wheel-shaped molecule B9 is aromatic like benzene. Because some of the compound’s delocalized electrons are σ electrons and have the same bonding pattern as the π electrons, the researchers conclude that the molecules exhibit σ and π aromaticity, so-called double aromaticity.

Other manifestations of nontraditional aromaticity include the σ aromatic PtZnH5 cluster reported by the University of California, Los Angeles’s Anastassia N. Alexandrova and coworkers, the δ aromatic Ta3O3 cluster reported by the Boldyrev and Wang team, and a 3-D aromatic fullerene analog B40 reported by Wang and other collaborators.

It’s true that none of these species have been “bottled.” But before Boldyrev and Wang got started, chemists had reported nontraditional compounds that had aromatic properties and could be isolated. For example, some researchers had been making metallabenzenes—predicted earlier by Hoffmann and his colleagues—in which a metal replaces a C–H group in the benzene ring. In addition, Gregory H. Robinson and coworkers at the University of Georgia provided the first solid evidence for all-metal aromaticity in 1995 when they prepared a gallium complex with a triangular Ga3 core.

“Aromaticity has in recent years become an important and powerful unifying concept to describe the stability and bonding in many chemical species beyond organic chemistry, including inorganic cations, anions, radicals, and clusters,” Boldyrev says. He doesn’t think the reactivity or the inability to bottle the species should preclude them from being labeled aromatic.

Hoffmann doesn’t deny that the molecules exist and can be studied. “But to label them as aromatic, with the 150-year-old history of thermodynamic stability, kinetic persistence, and chemical reactivity associated with that concept, as well as the 20th-century correlates we’ve added [for example, aromaticity in spherical molecules such as fullerenes], is to me—and I think to many—a stretch,” he says.

Hoffmann threw down a second gauntlet by offering “a good bottle of New York state Riesling” to anyone who prepares milligram quantities of the newfangled compounds. Hoffmann says he would even accept as convincing evidence a study on the dimerization or polymerization of one of the molecules or a study showing that it is persistent under ambient conditions.

Wang responded that a bottle of Riesling isn’t enough incentive to drop everything to make a vial of Na2Al4. He might do it for a bottle of champagne.

Gernot Frenking, a computational chemist at Philipps University, in Marburg, Germany, shares Hoffmann’s view on these new species. “There was a time when I was wondering if there were molecules left that do not exhibit some kind of aromaticity.” At the same time, Frenking doesn’t fault Boldyrev and Wang for picking up on the concept and showing that it can be useful for providing new understanding of the nature of matter.

“I personally do not care whether Al42– can be considered aromatic or not,” says theoretical chemist Dage Sundholm of the University of Helsinki, in Finland. “Aromaticity is just a name. What is more important to me is that molecules such as Al42– are formed in molecular beam experiments and how we can explain why this species is more abundant in the beam than other negatively charged aluminum clusters.

“I agree there is aromaticity hype,” Sundholm continues. “However, that will disappear in the future when we have invented better and more reliable methods to determine the degree of aromaticity, or whatever we agree to call it.”  

Read Boldyrev and Wang’s response to Hoffmann’s article, a rebuttal by Hoffmann, and perspectives from other researchers, then add your thoughts at



This article has been sent to the following recipient:

Raphael Berger (February 23, 2015 8:52 PM)
Classical organic aromatic compounds like PAHs form under non-oxidising high temperature conditions from organic materials, mimicking extraordinary stability. Anyway such compounds are still rich in energy and reactive. This can be already seen in their HOMO-LUMO gap which is much smaller than those we see in saturated compounds. This paradox is at the heart of the aromaticity concept.

Classical organic aromatics all show a strong diamagnetic response in the magnetic field. This is the so called "magnetic criterion" for aromaticity. The strong diamagnetic response can be readily explained in terms of (a) the comparably small HOMO-LUMO gap and (b) the symmetry of the frontier orbitals. It is relatively easy to see that the levels in these aromats are filled in way analogous to two-dimensional noble-gas atoms, which follow a 4n+2 aufbau.

In this way the magnetic criterion for aromaticity reflects this paradoxical situation between stability (filled molecular levels) and instability (small HOMO-LUMO gap). When transferred to entirely different chemical species like metal clusters or low-valent main group compounds situations may occur where there is a strong diamagnetic response but no signs of "stability" at the same time. In such cases talking about aromaticity and implying stability is an abuse of the underlying physical principles.

Robert Buntrock (March 13, 2015 3:59 PM)
Since Aromaticity dominated studies by Paul Schleyer the last decades of his life (cf. the interview of Paul by Steven Bachrach, author of Computational Organic Chemistry, 2nd ed.) he would have had much to say in this and other debates on aromaticity but for his death in January.
Robert Buntrock (March 13, 2015 4:04 PM)
(unsuccessfully attempted to post two weeks ago)
I think that Weinhold has the best take on this interesting controversy. The concept of aromaticity has evolved over the years but still lacks a succinct definition (per Schleyer et al. as recounted in Computational Organic Chemistry, 2nd ed., by S. Bachrach). Also not discussed is "classic" heterocyclic aromaticity, championed by A. Albert, which comes in two favors, pi-excessive and pi-deficient, which explains a lot of heterocyclic chemistry, both properties and reactions.
Alexander I. Boldyrev and Lai-Sheng Wang (March 20, 2015 3:41 PM)
When Kekulé introduced aromaticity in chemistry, he did not really know what was the reason for low reactivity and high stability of benzene and its derivatives. So, he used term aromaticity for unifying these chemical compounds since all of them had odor. Since then a lot of advancements have been made in order to explain those specific properties of benzene and its derivatives. Today we do believe that the main reason for the low reactivity of benzene is its multicenter nature of chemical bonding, which does not allow us to represent it by a single Lewis structure and forces us to use the resonance description with at least two Lewis (Kekulé) structures. The Hückel rules must also be obeyed for molecules to be aromatic. This delocalized bonding can now easily explain the low reactivity, high stability (resonance energy) and other specific properties of benzene. The presence of multicenter bonding was the driving force in extending aromaticity further into organic chemistry. We recognize today that the cyclopropenyl cation, C3H3+ is the simplest organic aromatic species; even it was not bottled yet. Similarly, we believe that the cyclopentadienyl anion, C5H5- is an aromatic species that is one of the most popular ligand in inorganic chemistry nowadays. π-bonding in both of these species also cannot be represented by a single Lewis structure and therefore bonding in those ions is also multicenter one. It would be too simplistic to think that multicenter bonding is a property of carbon only and a property of π-bonding. So, aromaticity, followed by the multicenter bonding, was extended into inorganic chemistry, to molecules containing main group elements and transition metals. It was also extended into σ-, δ-, and ϕ-bonding. Of course, aromaticity is a qualitative concept and if we want to quantify it, we need numerical descriptors such as resonance energy, NICS, HOMO, and other indices, which could give us more understanding on the importance of aromaticity in a particular chemical species. But aromaticity, similarly to chemical bonding, was and is a qualitative concept and that is why it is hard to characterize it specifically, very much like chemical bonding itself. As chemical bonding concept continues to concur new territories (solids, 2D-materials, nanoparticles, clusters), aromaticity follows the same trend.
Alexandru Balaban (April 6, 2015 4:31 PM)
Bench chemists’ search for similar features started when the strange chemical behavior of high C/H-ratio compounds related to benzene (favoring substitution, like alkanes with low C/H ratio) was contrasted with that of alkenes or alkynes (favoring addition). One feature (smell) led to the name “aromaticity,” but other, more relevant, features were soon discovered. Stepwise, starting with August Kekulé’s formula, theoretical contributions by Erich Hückel, Linus Pauling, Robert Robinson, Michael Dewar, and Eric Clar followed, shedding light on the importance of π-orbital overlap in coplanar structures. Sason Shaik and Philippe Hiberty evaluated the relative contributions of σ and π electrons to the bonding. As the number of chemical structures in the Chemical Abstracts database increased reaching today’s value of almost 70 million, it became apparent that heterocyclic aromatics were the majority among the organic compounds, which in turn are the vast majority of “bottleable” chemical substances.

The extra stability due to cyclic conjugation is the most striking feature of aromaticity, and this feature goes hand-in-hand with the preference of bench chemists for defining aromatics as “bottleable” compounds. High stability can be too much of a good thing, as shown by the formation of carcinogenic compounds in combustion, and by the persistence of DDT, PCB, or TCDD. Theoretical estimation of bond lengths (which gradually become more easily accessible experimentally) or of conjugation energies can serve to establish relative scales of aromaticity for substances, or of local aromaticity in rings of polycyclic benzenoids. An overall degree of (anti)aromaticity can be determined both experimentally and theoretically by the diamagnetic ring current in aromatics, and the paramagnetic ring current in antiaromatics. Paul Schleyer’s theoretical prediction called nucleus-independent chemical shift (NICS) at various distances from the molecular plane is useful, but because π-electron systems exist which sustain diamagnetic currents in some rings and at the same time paramagnetic ring currents in others, aromaticity cannot be defined in terms of ring current for polycyclic systems, as argued by D. E. Jung (Grenzen der Ringstromdefinition der Aromatizität: Ringstromberechnungen in nicht-alternierenden Tri- und Tetracyclen, Tetrahedron 1969, 25, 129–134).

Having studied in depth pyrylium salts, which have the conjugated six-membered aromatic ring with the strongest possible single-atom perturbation by their very electronegative oxygen heteroatom, I am familiar with substances that are weakly aromatic, do not undergo substitutions, ring-open and ring-close easily, so that they are valued synthons for making a large variety of structures with higher aromaticity. An interesting fact is that both tropone, and the 4H- or 2H-pyrones, have zero aromatic character according to practically all theoretical and experimental criteria, but they do become aromatic (as hydroxyl derivatives) at lower pH values. This fact presents analogies to discontinuities in mathematical functions.

A remarkable feature of aromatic rings is their ability to stabilize chemical structures, a feature that has not been mentioned so far. Familiar examples are triarylmethyl free radicals or cations, diazonium cations, phenoxide anions, arylpentazoles, and azobenzene contrasted with unsubstituted or alkyl-substituted analogs. The ease and stereoselectivity of pericyclic reactions were explained by the Woodward-Hoffmann Rules. One should not neglect the “click reactions” yielding aromatic triazoles by a Huisgen cycloaddition involving alkynes and azides, which has become a highly-valued synthetic tool also for medicinal chemists.

It is remarkable that in the 1960s the trend was either to multiply various restricted hyphenated aromaticities (such as quasi-, pseudo-, homo-, non-, anti-) or even to discontinue using this “ill-defined fuzzy concept.” By contrast, at present there is the tendency to widen the borders of aromaticity in order to embrace also fleeting short-lived molecules, represented by Boldyrev and Wang. There is some merit in including 3-D covalent structures such as nanotubes and fullerenes among aromatics, because they contain benzenoid rings. However, I would not consider carboranes with their three-center bonds as aromatic, although they are “bottleable” and undergo electrophilic substitution reactions.

One can draw a parallel with the notion of an element. Chemists have accepted as “new elements” the few atoms with extremely short half-lives obtained by nuclear reactions, if there is proof about the atomic number of these atoms, so that the “bottleable criterion” is no longer a universal requirement for elements, as it was in 1900. Two years were spent by Pierre and Marie Curie for obtaining a bottleable amount of radium chloride to validate their discovery of this element that, unlike the highly-radioactive polonium-209 (with T½ = 102 years, which they had first discovered), is sixteen times less radioactive (T½ = 1600 years for radium-226).

However, probably like most bench chemists, I would agree with Roald Hoffmann about the reluctance to include as aromatic the substances described by Boldyrev and Wang. I recall that a similar controversy between Paul Schleyer and Alan Katritzky about the criteria for aromaticity ended with the agreement that aromaticity is a multi-dimensional concept (M. K. Cyranski, T. M. Krygowski, A. R. Katritzky, P. v. R. Schleyer, J. Org. Chem. 2002, 67, 1333). Is there a similar middle ground in the present discussion?

• A.T. Balaban, Is aromaticity outmoded? Pure Appl. Chem., 1980, 52, 1409.
• A.T. Balaban, D.C. Oniciu, and A.R.Katritzky, Aromaticity as a cornerstone in heterocyclic chemistry. Chem. Rev. 2004, 104, 2777.
• A.T. Balaban and T. S. Balaban, Pyrylium Salts (Update 2013), Science of Synthesis. Knowledge Updates 2013/3, G. Thieme Verlagg, Stuttgart, 2013, pp. 145-216.
• A.T. Balaban and M. Randić, Structural Approach to Aromaticity and Local Aromaticity in Conjugated Polycyclic Systems, In: Carbon Bonding and Structures: Advances in Physics and Chemistry. Carbon Materials: Chemistry and Physics 5 (M. V. Putz, ed.), Springer, Berlin, 2011, pp. 159-204.
• A. T. Balaban, Monocyclic hetarenes with π-electron aromatic sextet, Advances in Heterocyclic Chemistry (ed. A. R. Katritzky), Vol. 99, pp. 61-105, Elsevier, Amsterdam, 2010.
• A. T. Balaban, Ode to the chemical element carbon. In: Exotic Properties of Carbon nanomatter: Advances in Physics and Chemistry (eds. M. V. Putz and O. Ori), Springer, Dordrecht, 2015, pp. 1-18.
Harold Teague (May 11, 2015 11:22 AM)
Until enhanced stability due to ring circulation (of pi electrons) is demonstrated by NMR or a similar method, I prefer a term other than aromaticity.

Leave A Comment

*Required to comment