Platinum(II) Di-ω-alkenyl Complexes as "Slow-Release" Precatalysts for Heat-Triggered Olefin Hydrosilylation

We describe the synthesis, characterization, and catalytic hydrosilylation activity of platinum(II) di-ω-alkenyl compounds of stoichiometry PtR₂, where R = CH₂SiMe₂(vinyl) (1) or CH₂SiMe₂(allyl) (2), and their adducts with 1,5-cyclooctadiene (COD), dibenzo[a,e]cyclooctatetraene (DBCOT), and norbornadiene (NBD), which can be considered as slow-release sources of the reactive compounds 1 and 2. At loadings of 0.5 × 10^-6-5 × 10^-6 mol %, 1-COD is an active hydrosilylation catalyst that exhibits heat-triggered latency: no hydrosilylation activity occurs toward many olefin substrates even after several hours at 20 °C, but turnover numbers as high as 200000 are seen after 4 h at 50 °C, with excellent selectivity for formation of the anti-Markovnikov product. Activation of the Pt^II precatalyst occurs via three steps: slow dissociation of COD from 1-COD to form 1, rapid reaction of 1 with silane, and elimination of both ω-alkenyl ligands to form Pt⁰ species. The latent catalytic behavior, the high turnover number, and the high anti-Markovnikov selectivity are a result of the slow release of 1 from 1-COD at room temperature, so that the concentration of Pt⁰ during the initial stages of the catalysis is negligible. As a result, formation of colloidal Pt, which is known to cause side reactions, is minimized, and the amounts of side products are very small and comparable to those seen for platinum(0) carbene catalysts. The latent reaction kinetics and high turnover numbers seen for 1-COD after thermal triggering make this compound a potentially useful precatalyst for injection molding or solvent-free hydrosilylation applications.

Dynamics of the efficient cyclometalation of the undercoordinated organoplatinum complex [Pt(COD)(neoPh)]+ (neoPh = 2-methyl-2-phenylpropyl)

The cyclometalation reaction of [Pt(COD)(κ¹-neoPh)]⁺ (neoPh = 2-methyl-2-phenylpropyl) to [Pt(COD)(κ²-neoPh)] was studied experimentally and mechanistically using DFT and MD simulations.