Triple bonds take on a very different shape in heavier atoms where relativistic quantum effects are at play, researchers in the US have found. The researchers say their spectroscopic study on the triple-bonded CBi– anion shows how molecular orbitals mix together, as predicted by theory, but also reveals unexpected differences that raise important questions about bonding in heavier atoms.
The conventional understanding of chemical bonding relies on the notion that valence electron orbitals on bonding atoms overlap with each other to form the bonds between the atoms. The atomic s or p orbitals combine in a linear manner to form σ and π orbitals, for example, creating an energy ladder that fills up as extra electrons are added. In this picture, a triple bond is made up of a σ orbital and two π orbitals, each occupied by a pair of electrons with opposite spin. But this only works when the angular momentum of the orbital – which controls its shape – and electron spin are treated separately. When electrons are moving at close to the speed of light, this model breaks down because those two quantities mix – a phenomenon known as spin–orbit coupling – and σ and π orbitals with the same total angular momentum can combine with each other. This becomes increasingly relevant in the heavier elements where orbitals contract under the influence of the greater nuclear charge and so their electrons move faster.
Molecular orbitals get mixed up
Physical chemist Lai-Sheng Wang and his group at Brown University used cryogenically cooled, angle-resolved photoelectron spectroscopy to study the shape of the electronic orbitals in CBi–.1 By irradiating the anion with photons and then measuring the electrons that are subsequently ejected, they reconstructed the average electron density in the molecule to create images of its molecular orbitals. As theory predicts, they saw some mixing of the σ and π orbitals, but Wang’s student Deniz Kahraman was surprised to find that the two π orbitals – which he expected to appear identical – looked completely different from each other. ‘Deniz said he couldn’t find any evidence in the literature that this was known,’ recalls Wang. ‘I said: “If nobody has observed this … it’s really important because it’s a fundamental question”.’ Further experimental work with another member of the group, Jie Hui, confirmed what Kahraman had observed. ‘At this point, I decided I didn’t know what we were talking about,’ says Wang.
Wang consulted an old friend, theoretician Kirk Peterson at Washington State University. Peterson calculated that the difference came from the differing angular momentum values of each π orbital. The π orbital with angular momentum ½ is able to mix with the σ orbital -which also has total angular momentum ½, producing two bonds with both σ and π character. The π orbital with angular momentum ³⁄₂ was not able to mix, as this would not conserve total electronic angular momentum. ‘I did some calculations and it’s only about 20% of the σ character that mixes into [the π-½], but that seems to be enough to change the character – at least visually – of the electron distribution so it looks very σ-like,’ says Peterson. The researchers believe this change warrants further investigation. They also intend to study progressively lighter elements in the same group to ascertain the point at which, in Wang’s words ‘the π-½ actually behaves like a π’, and to ascertain whether empty σ orbitals affect π-orbital splitting, as theory predicts they should.
Quantum chemist Pekka Pyykkö at the University of Helsinki in Finland says the study is ‘new and original’. ‘It is very clean work, with the best available methods,’ he says. He is less surprised by the results than the authors, however: ‘The experts have been expecting something like this all the time,’ he says. ‘It all goes back to Kenneth S Pitzer, who has been [cited] here.’ As far back as the 1970s, he says, Pitzer hypothesised that effects such as the weakening of covalent bonds in thallium compounds could be connected to spin–orbit coupling ‘messing up covalent bonding’.2
