At the beginning of the project, it was unknown whether alpha-bromide-silylethers (A, Figure 2) would be applicable with Ni-catalysis. It was clear from previous reports that the regioselectivity was highly dependent on the substrate substitution pattern. Therefore, to avoid such substrate control, an alternative precursor class (B, Figure 2) was proposed containing a phthalimide moiety that should also undergo reductive activation but in this case generating an electrophilic oxy-radical intermediate with an alternative reactivity profile. Precursors of this structure were unknown so an efficient synthesis was devised. Reaction of the alcohol and N-hydroxyphthalimide across a dichlorosilane was realized after careful selection of the reaction conditions to avoid the formation of side-products. Although challenging to purify, B proved stable on the bench under air allowing investigation into its reactivity. Unfortunately, no positive results were obtained for the proposed remote carboxylation reactions (Figure 2), nor any other catalytic C–C bond forming reactions.
See Figure 2
It was eventually discovered during the investigation of hydroxyl-silylethers as a possible precursor class (C, Figure 3, top) that the generation of the desired silyoxy-radical leads to the formation of alkyl radicals via intermolecular addition onto another silyl-center, likely driven by the formation of the strong Si−O bond. Consequently, this suggested that B or C are not suitable precursors to direct remote radical functionalizations of alcohols. On the other hand, the possibility to catalytically generate alkyl radicals from silanols is in itself attractive and was explored following this discovery (Figure 3, bottom). The generality of this protocol is now under investigation.
See Figure 3
The utilization of α-bromo-silylether precursors (A, Figure 2) yielded more positive results. Although no reactivity for the proposed remote carboxylation was detected, an alternative remote arylation reaction was possible (Figure 4) with selectivity for the γ-position observed. Side-reactions such as arylation and homocoupling at the initial α-silylether position were significant and extremely challenging to avoid. Consequently, a maximum yield ~30% was achieved for this remote-arylation protocol employing both heterogeneous metal reductant (Figure 4, left) or photoredox catalysis conditions (right), even after extensive study of all reaction variables.
See Figure 4
The proposed silyl-directed carboxylation methodology may have been extended to include amines in which β-carboxylation would generate β- or γ-amino acids (Figure 5, top), valuable building blocks that have shown promise as peptidomimetics. However, α-bromo-silylamines (D) are even less stable than their alcohol analogues. An alternative strategy could be to apply aziridines in which Ni-catalyzed C(sp3)-N cleavage followed by CO2 insertion would also lead to beta-amino acid products. This proved possible with the application of hindered and electron rich bipyridine ligand L key for high selectivity in the Ni-catalyzed carboxylation of N-Tosyl aziridines under CO2 atmoshphere (Figure 5, bottom). Most intriguing about this protocol was the almost unique ability of MeOH to promote reactivity, with mechanistic studies pointing towards it playing a role in stabilizing key Ni-species such as the ring-opened zwitterionic intermediate proposed by Hillhouse. As well as the high impact publication that resulted from this work, it was disseminated by oral presentation at a Merck Sharp & Dohme–ICIQ networking event.
See Figure 5
