When ligands bind to a transition metal its valence d shell gets split, which has interesting consequences if the shell is partially filled. Here, the energy differences between frontier orbitals can be small enough to make the metal a good σ-acid and π-base — useful properties when it comes to binding substrates such as H2 and CO. Chemists seek to replicate this behaviour using low-valent compounds of main-group elements, yet these species rarely have the appropriate atomic or electronic structure to afford stable adducts. A team led by Stephan Schulz and Peter Schreiner have hit back by preparing a silylene that enjoys steric and electronic protection in the form of bulky Ga-donor ligands. This privileged silylene, which the team now describes in Nature Chemistry, is ideally set up to split H2 or form a rare complex with CO.

Carbenes are likened to transition metals because C has two valence electrons to distribute between an sp2 hybrid orbital and a higher-lying p orbital. A common strategy to stabilize a carbene is to introduce π-donor substituents that increase the singlet–triplet energy gap, but this often robs the carbene of its ability to activate certain bonds. Enter the silylenes — divalent Si analogues of carbenes that exist as singlet species yet have distinct reactivity.

Credit: David Schilter/Springer Nature Limited

Early matrix-isolation work on the simplest silylene, H2Si:, uncovered a non-planar H2Si:–CO complex in which Si plays the synergistic role of σ-acid and π-base reminiscent of transition metals. Schulz, Schreiner and their colleagues used density functional theory calculations to show that this complex is 16 kcal mol−1 lower in energy than the silaketene H2Si=CO. By contrast, H2C: has greater access to its triplet state, such that its singly occupied sp2 and p orbitals engage in typical σ and π covalent bonding with CO to give ketene H2C=CO, which now lies 6 kcal mol−1 lower than the non-planar isomer H2C:–CO.

Despite its fragility, the observation of H2Si:–CO inspired the team to prepare a more stable analogue so they could probe its bonding and chemistry. For this they turned to other main-group chemistry they had developed, namely insertion reactions of unsaturated Ga species LGa (L− = N,Nʹ-bis (2,6-diisopropylphenyl)-1,3-pentanediiminate). Thus, they converted SiBr4 to [L(Br)Ga]2SiBr2, a tetravalent species that undergoes debromination when treated with a further equivalent of LGa to generate a putative [L(Br)Ga]2Si: intermediate. If left to its own devices, the Si centre inserts into a C–C bond of its own ligand, so Schulz, Schreiner and colleagues conducted debromination under CO in hopes of trapping it as a complex (when under H2 the reaction affords silane [L(Br)Ga]2SiH2). Gratifyingly, the team isolated [L(Br)Ga]2Si:–CO and its X-ray structure reveals a non-planar arrangement that enables pSi←spC σ-bonding and synergistic sp2Si→π*CO backbonding. But the same arrangement exists in H2Si:–CO, so why is [L(Br)Ga]2Si:–CO stable at room temperature? Calculations on [L(Br)Ga]2Si: indicate that the bulkiness of the L(Br)Ga groups brings them apart and raises the energy of the sp2Si lone pair, which is now a better π-donor. Conversely, overlap of vacant p orbitals on the Si and Ga centres ensures that the pSi orbital is a good σ-acceptor. Curiously, π-bonding dominates, so it is arguable if we can even call it ‘backbonding’. “It took us a while to get our head around this but it is what makes bonding possible in the first place,” notes Schreiner.

It took us a while to get our head around this but it is what makes bonding possible in the first place

Like many a metal carbonyl, [L(Br)Ga]2Si:–CO undergoes substitution with isonitrile CNCy to give [L(Br)Ga]2Si:–CNCy. “I am pretty sure that the mechanism is associative because it proceeds at 25 °C while the precursor is stable to decarbonylation up to 80 °C,” predicts Schulz. The team is now investigating ligand exchange and Si-assisted reactions at the CO ligand itself. This chemistry is a hallmark of transition metals and will no doubt motivate more investigations into such fertile main-group systems.