Potential Carbon–Fluorine Reductive Elimination from Pincer-Supported Rh(III) and Dominating Side Reactions: Theoretical and Experimental Examination

Abstract

This paper explores the potential for C–F reductive elimination from Rh(III) pincer complexes. A DFT computational study indicated that concerted C–F reductive elimination from (POCOP)Rh(CHCH2)(F) (3) (where POCOP is κ3P,C,P-2,6(iPr2PO)2C6H3, and aryl/bis(phosphinite) pincer ligand) possesses an experimentally plausible activation barrier of ΔG⧧ = 28.3 kcal/mol. This barrier is considerably lower than that calculated (35.7 kcal/mol) for the analogous C–F reductive elimination from (POCOP)Rh(Ph)(F) (1). The difference is ascribed to the need for a partial rotation of a phenyl or vinyl group in the transition state, with the phenyl being more encumbered by the steric bulk of the supporting pincer ligand. DFT calculations did not analyze the full range of possible side reactions, which have proven to be dominant. The attempted synthesis of 1 was unsuccessful because of competing C–C reductive elimination at the stage of the preparation of the (POCOP)Rh(CHCH2)(I) precursor. DFT calculations predicted C–C reductive elimination to be facile but markedly unfavorable in the monomeric unit because of inherent strain in the product. That strain is apparently relieved in dimerization that takes place with opening up of the pincer to become bridging between two Rh centers. The opening up of the pincer, and thus dimerization and C–C reductive elimination, was prevented by the use of the tBuPOCOP ligand (κ3P,C,P-2,6(iPr2PO)2-3,5-But2C6H3), which allowed isolation of (tBuPOCOP)Rh(CHCH2)(I) (11). Compound 11 was converted to (tBuPOCOP)Rh(CHCH2)(OTf) (12) via reaction with AgOTf. However, treatment of 11 with AgF or of 12 with CsF failed to produce (tBuPOCOP)Rh(CHCH2)(F), resulting instead in multiple products containing P–F bonds. On the other hand, 12 was cleanly converted to (tBuPOCOP)Rh(CHCH2)(OBut) (13) via reaction with NaOBut and then to (tBuPOCOP)Rh(CHCH2)(OC6H4F-p) (14) by treatment of 13 with p-fluorophenol. Neither 13 nor 14 gave any evidence of C–O reductive elimination. Instead, thermolysis of either 13 or 14 resulted in dehydroalkoxylation to the bimetallic divinylacetylene complex (tBuPOCOP)Rh(CH2═CHC≡CCH═CH2)Rh(tBuPOCOP) (15) as the major product.