You spin me right round - Computational and experimental studies on conformer-specific reactivity in hydroxycarbenes

Abstract

Hydroxycarbenes (R–C̈–OH) are known ligands in so-called Fischer-complexes. Still, the scientific community debated the existence of “free” hydroxycarbenes until the isolation of the simplest hydroxycarbene, hydroxymethylene (R=H), which Schreiner et al. achieved at cryogenic temperatures. Surprisingly, hydroxymethylene is not persistent and reacts quickly to formaldehyde despite a high barrier, concluding that quantum mechanical tunneling (QMT) must be involved. As QMT is not only dependent on the reaction barrier height but also on the reaction barrier width—which hard to model—there is a significant challenge in the rational design of reactions. Especially, the reduction of a conformer's activation barrier to its product has a distinct influence on the product ratio (Curtin-Hammett principle). Nevertheless, we have not observed the s-cis-conformer of hydroxycarbenes and their reactivity. Hence, we explored the generation of s-cis hydroxycarbenes by applying two different strategies. In the first publication, we tried to isolate the s-cis-conformer of trifluoromethylhydroxycarbene (F₃C–C̈–OH), which is stabilized by an intramolecular H–F bridge. Indeed, after irradiation with UV light, we could observe the s-cis conformer. The s-trans-conformer displays the typical [1,2]H-shift to trifluoroacetaldeyde through QMT, whereas the s-cis-conformer does not transform. Hence, trifluoromethylhydroxycarbene shows conformer-specific tunneling. The established Curtin Hammett principle is not applicable because a significant activation barrier separates the two conformers, making fast equilibration impossible. At thetime of the publication, trifluoromethylhydroxycarbene showed the longest measurable half-life (𝜏 = 144 h) of all previously isolated hydroxycarbenes. Hence, the push-pull substitution of the substituents OH and CF₃ may stabilize carbenes, which was later observed for other hydroxycarbenes. In the second publication, we addressed whether it was possible to generate conformers in hydroxycarbenes through stimulation with near-infrared light (NIR). We investigated the system of dihydroxycarbene, which has three different conformers and is an essential intermediate on the [H₂CO₂] potential energy surface. The s-cis,s-trans-conformer and the s-trans,s-trans-conformer were known, but not the s-cis,s-cis-conformer, an essential intermediate in the bimolecular reaction of H₂ and CO₂. We could transform either conformer to the previously unobserved s-cis,s-cis conformer through NIR irradiation. At 3 K, the s-cis,s-cis-conformer is not stable and undergoes QMT back to the s-cis,s-trans-conformer within 22 minutes. We found indications for a second QMT reaction of the s-cis,s-cis-conformer, possibly a decomposition to CO₂ and H₂, which we could confirm by computations. In the third publication, we presented a program suite to compute tunneling half-lives in a simple fashion (TUNNEX) using the Wentzel, Kramers, and Brillouin (WKB) method. The necessity arose during the last project as to accelerate the computation of QMT half-lives. After generating the reaction path using Gaussian—a widely used quantum chemistry program—the user deploys a helper tool to generate the data necessary for WKB computation. After import in TUNNEX, the user defines the parameters and computes the QMT half-life. The program is free to use and published online. It uses openly available libraries for interpolation and integration of the tunneling path. TUNNEX is the first entry point for computing the QMT half-lives and is as user-friendly as possible.