Catalyst activation and the dimerization energy of alkylaluminium compounds

Abstract

Alkylation of olefin polymerization catalyst precursors with Me3Al (or MAO) is a crucial first step in the activation of many polymerization catalysts. However, structures with alkyls and/or halides bridging between aluminium atoms, are known problem cases for DFT methods. We have explored the scope of the problem and possible solutions. The performance of various DFT methods for a range of bridged dimers [H2M]2(μ-X)2 and [Me2M]2(μ-X)2 (M = Al–Tl, X = H, Me, NMe2, OMe, F, Cl) and related compounds was investigated in a systematic fashion. In the absence of reliable experimental data, highly accurate calculated dimerization and complexation energies (extrapolated to the CCSD(T,Full)/aug-cc-pVQZ level of theory: fully correlated Coupled-Cluster Theory with Single and Double and Perturbative Triple excitations) for 26 model systems relevant to Ziegler–Natta catalysis and related chemistry serve as reference. Most functionals (strongly) underestimate the dimerization energy for both electron-deficient (hydride, alkyl) and regular (halide, alkoxide, amide) bridges, and for both main-group and transition metals. To reach “chemical accuracy”, a basis set of VTZ quality (with BSSE correction) or VQZ (without BSSE correction) is recommended, in combination with the M06-2X functional, or with the TPSSTPSS functional plus Grimme's D2 correction. Subsequently, we explored alkylation of olefin polymerization catalyst precursors LTiX2 (L = bis(cyclopentadienyl), Salan or Salalen) with Me3Al to LTiMe2. Using our recommended procedure, we find that the process is exergonic for most electronegative groups X (F, Cl, OR, NR2) provided the dimerization energies of Me3Al and Me2AlX are taken into account. Moreover, alkylation is more exergonic for metallocenes, which we ascribe to the more saturated nature of the Cp2M fragment. Alkylation is predicted to be more favourable for amides and alkoxides than for the widely used chloride precursors, and failure to exploit those precursors in catalysis is likely due to kinetic issues caused by steric factors. In particular, employing metal alkoxides or amides with small alkyl groups (OMe, OEt) rather than the commonly used bulky OiPr and OtBu groups might yield catalyst precursors that can be activated more easily than chlorides.