Henry Rzepa's blog

The two previous posts have explored one of the oldest bonding rules (pre-dating quantum mechanics), which postulated that filled valence shells in atoms forming molecules follow the magic numbers 2, 8, 18 and 32. Of the 59,025,533 molecules documented at the instant I write this post, only one example is claimed for the 32-electron class. Here I suggest another, Nobelocene (one which given the radioactive instability of nobelium, is unlikely to be ever confirmed experimentally!)

Nobelium has the electronic configuration [Rn].5f¹⁴.7s² , which means the 6d and 7p shells are still empty. Filling these would take 10+6 electrons, or four more electrons (20) if one starts from No⁴⁺, resulting in a complete 32-electron filled shell for the Nobelium. These twenty electrons could be provided by two cyclo-octatetraenyl (COT) dianion ligands. Nobelium, with its nuclear charge of +102, has highly relativistic inner-shell electrons, and so special techniques must be used to model this. Here I have used a SARC all electron relativistically contracted basis set (DOI: 10.1021/ct100736b), to be used with the Douglas−Kroll−Hess scalar relativistic Hamiltonian (for details, see here). The QTAIM analysis is shown below (quite a spider’s web):

There are 16 bond critical points located along the lines of each No..C, with ρ(r) 0.03. This is a very low value indeed for a covalent bond, being of the same order as strong hydrogen bonds, and so should be classed as an interaction rather than a bond. ELF basins cannot normally be located for hydrogen bonds, and neither can they here. Nobelocene in this regard is pretty boring, being almost entirely ionic.

Rather more interesting are the molecular orbitals. The most stable π-type is shown below. Many of the orbitals show the Nobelium atomic orbitals non-interacting with the ligand, probably because the relativistic contraction renders them inert to mixing, in a manner which is often used to explain the inert nature of e.g. the Pb 6s² electrons in divalent lead.