The cationic rhodium(I) complexes [Rh(1,5-cod)2]A (A = OTf, PF6, 1,5-cod = 1,5-cyclooctadiene) have been shown to exhibit high catalytic activity for the transition metal-mediated ring-opening polymerization (ROP) of silicon-bridged [1]ferrocenophanes (e.g.,
fcSiMe2 (1a), fc = Fe(η5-C5H4)2), to quantitatively afford high molecular weight (M
n > 105)
polyferrocenylsilanes (e.g., [fcSiMe2]
n
(2a)). However, prolonged exposure of the resulting
polyferrocene solution to the catalyst resulted in a dramatic molecular weight decline.
Transition metal-catalyzed ROP of tetramethyldisilacyclobutane (6) and copolymerization
of this species with 1a also proved readily feasible with [Rh(1,5-cod)2]OTf forming high
molecular weight polycarbosilane homopolymer (7) (M
n = 4.20 × 105 to 1.08 × 106, PDI =
1.39−2.08) and poly(carbosilane-r-ferrocenylsilane) (8) copolymer (M
n = 4.63 × 105, PDI =
1.86), respectively. While [Rh(1,5-cod)(dmpe)]PF6 (dmpe = bis(dimethylphosphinoethane))
proved less active as a ROP catalyst, this species represents the first phosphine-ligated
catalyst for the ROP of 1a, and moreover, in this case no subsequent depolymerization of
the resulting poly(ferrocenyldimethylsilane) (2a) was detected. The cleavage of the polyferrocene backbone in the presence of [Rh(1,5-cod)2]OTf was modeled through the catalytic
reaction of Fc2SiMe2 (Fc = Fe(η5-C5H5)(η5-C5H4)) with the rhodium complex (9 mol %), which
resulted in Cp−Si bond cleavage affording ferrocene as the sole, isolable product. The
presence of 1,5-cod inhibited the catalytic activity of both [Rh(1,5-cod)2]OTf and [Rh(1,5-cod)(dmpe)]PF6, suggesting that elimination of a 1,5-cod ligand is necessary to generate the
true active catalytic species. The addition of dmpe effectively arrested the catalytic activity
with [Rh(1,5-cod)(dmpe)]PF6; this is presumably a result of the displacement of the 1,5-cod
ligand by dmpe, which would generate catalytically inactive [Rh(dmpe)2]PF6.