Density functional benchmarks for methylaluminoxane: successes, biases, and transferability limits.

Area of Science:

• Computational chemistry
• Catalysis science
• Polymer chemistry

Background:

• Methylaluminoxane (MAO) is a vital cocatalyst in olefin polymerization.
• The precise molecular structure and reactivity of MAO are not fully understood.
• Existing computational methods for MAO lack sufficient benchmarking, hindering accurate predictions.

Purpose of the Study:

• To comprehensively benchmark 25 density functional methods for their accuracy in describing methylaluminoxane (MAO) oligomers.
• To evaluate the performance of these methods for geometry, electronic energies, vibrational frequencies, and thermodynamic properties.
• To identify reliable computational approaches for studying MAO structure and reactivity in olefin polymerization.

Main Methods:

• Benchmarking 25 density functional theory (DFT) methods against high-level DLPNO-CCSD(T) and RI-MP2 reference calculations.
• Evaluation of MAO oligomers, focusing on geometry, electronic energies, vibrational frequencies, and thermodynamic properties.
• Analysis of energetic biases and structural preferences of different DFT functionals.

Main Results:

• Ten DFT functionals exhibit significant energetic biases towards specific MAO structural motifs (µ4-O and µ-Me), with errors up to 45 kJ mol⁻¹.
• ωB97X-D4 and MN15 functionals show the largest errors and are unsuitable for pure MAO systems.
• ωB97X-V and ωB97M-D4 functionals accurately reproduce electronic energies without structural bias for MAO oligomers.
• Quasi-harmonic treatment effectively minimizes method-dependent entropy variations for vibrational properties.
• Extending calculations to metallocene-MAO ion pairs revealed system-dependent functional behavior.

Conclusions:

• Functional performance in computational studies of MAO is highly system-dependent.
• ωB97X-V and ωB97M-D4 are recommended for accurate electronic energy calculations of MAO oligomers.
• Further investigation is needed for metallocene-MAO systems due to observed distinct behaviors.
• These findings have significant implications for selecting appropriate computational methods to study catalyst activation mechanisms in olefin polymerization.