# Is a compound aromatic if it also has anti-aromatic rings?

If I use Hückel's rule for the whole molecule, I get the answer as anti-aromatic. But the compound has two benzene rings which are individually aromatic.

• I'd go as far as to say there is no such thing as aromatic compounds. Sure, people say it all the time, and we even have a tag to that effect; so what? This is just a shorthand for "compounds with aromatic rings". Aromaticity is a property of one ring, not of the whole compound. A molecule, as you know, may be arbitrarily complicated. It may involve some anti-aromatic rings together with some aromatic ones. Then you'll be compelled to consider such compound both aromatic and anti-aromatic at the same time, which does not feel well. – Ivan Neretin Aug 20 '16 at 7:01

The molecule actually optimizes to a planar, delocalized $\pi$ system

Using B3LYP/6-31G(d) optimizations, we see that the molecule optimizes to become planar. (The initial configuration was intentionally bent.)

Moreover, if I plot the HOMO, it's highly delocalized:

Similarly, the LUMO is also delocalized across the entire molecule:

Consequently, I don't see any indication that the "side" benzene rings are localized from the cycloocta-1,5-diyne ring. The bond lengths are in fairly good agreement with the crystal structure:

Here's the crystal structure from J. Am. Chem. Soc., 1975, 97 (3), 658–659. Note there is a profound triple bond character, but the bridging bonds (1.44 Å) are somewhere between a typical aromatic bond length (1.40 Å) and a full single bond (~1.54 Å).

• I will admit my initial intuition was that the central ring was non-planar, but figured "why not run a calculation and see?" – Geoff Hutchison Aug 23 '16 at 18:12
• That is bloody weird. Glad you ran the calc. Would you add HOMO/LUMO +/- 3 energies to your post? – Lighthart Aug 23 '16 at 18:25
• Guys, guys, that's how stabilisation via anellation with aromatic rings works. Benzene rings are destabilised, but central more stable. Cyclooctadienediyn most probably doesn't exist. – Mithoron Aug 23 '16 at 18:46
• @Mithoron - yes, exactly. That's why I wanted to add a better answer for this one. – Geoff Hutchison Aug 23 '16 at 18:50
• @Mithoron Cyclooctadienediyn could exist (at least in silico) and I would like to see the contributions to the stabilization. Dissected into destabilization of benzenes by geometric deformation and also some Huckel-like diagram. Could be a nice paper in J. Chem. Education ;) – ssavec Aug 24 '16 at 18:50