AmericanChemicalSociety.com

Methylaluminoxane (MAO) is a commonly-used activator for olefin polymerization catalysts. MAO is made by hydrolysis of trimethylaluminum. The result of the hydrolysis is a poorly-defined oligomeric material that has defied definitive characterization for decades. We've developed two key methodologies that allow us to probe MAO in detail: (1) anaerobic electrospray ionization mass spectrometry (ESI-MS), achieved through the coupling of a glovebox to the instrument and (2) facilitation of the analysis of non-polar solvents using the same technique through the addition of ionic adulterants.

Early in the project, we discovered the need for further rigor in our analysis: the pyrophoric nature of the reagent has led to the requirement of further modification of our methods: ESI sources operate at negative pressure and are not sealed sufficiently well; we've modified the source to operate at positive pressure (set at +3 mbar) by adding an extra N₂ source and pressure gauge. This additional precaution has allowed us to reliably and reproducibly probe the speciation of MAO. The mass spectrum of commercial MAO in non-polar solvents is extremely weak in the absence of ionic adulterants, but analysis in their presence allows for the detection of a host of oligomeric species. No significant species are noted below 1000 m/z, but the region from 1200 – 2700 m/z consists of a rich suite of species that can be assigned to an unambiguous composition using a combination of m/z value, isotope pattern and chemical reasonableness (i.e. the charges must balance). Approximately 20 species appear with an intensity of >15%, whose reproducibility is excellent and whose identity has been established through the methods described above. Many less intense signals also appear (>100), and MS/MS data is being collected on these where practicable.

While mass spectrometry can be used to establish compositional information, e.g. [Me_xAl_yO_z], in the case of MAO complications arise through the equivalence of, for example, Al₂O₃ + AlMe₃ = (MeAlO)₃. The diagnostic tool in this case is MS/MS, and particularly noteworthy is the favoured fragmentation pathway of loss of AlMe₃. This behavior is entirely reasonable; it is a stable molecule in its own right and its removal from the oligomer involves the breaking of a single O®Al dative bond. Removal of more integrated fragments such as MeAlO and Al₂O₃ is observed, but usually at higher values of collision energy. We count the number of AlMe₃ fragments and from this data establish the ratio of Al₂O₃ : AlMe₃ : MeAlO. The amount of AlMe₃ typically represents 20-40% of the aluminum in each oligomer, a small core of Al_(2n+1)O_(3n+2) (n = 0-4) is a feature of most species and the remainder of each structure is made up of oligomeric (MeAlO). The charge arises when the oligomer acquires an anion (X^–), so the mass spectrometrically-observed distribution represents the most strongly Lewis-acidic components of the mixture.

The role of the MAO includes scavenging catalytic mixtures for residual moisture, and we can observe this behavior when less rigorous conditions are employed: the average Me / O ratio drops through the conversion of AlMe₃ to MeAlO and the appearance of additional Al₂O₃. MAO is also used to transform transition metal halide precursors such as Cp₂ZrCl₂ to the dimethyl analogue, and investigation of this reaction revealed the expected appearance of chlorine-containing isotope patterns in the mass spectrum, and their appearance at m/z values in increments of 20 m/z higher with each addition of Cl (which has a diagnostic isotopic signature compared to the predominantly monoisotopic C, H, Al and O) point to replacement of Me with Cl. The MS/MS of these species indicates that the location of the Cl is in place of Me on the AlMe₃, as AlMe₂Cl fragments are lost from the precursor ion in place of AlMe₃.

With compositional (and, thanks to MS/MS, some structural) information in hand, we are just beginning to apply theoretical methods to the knottier problem of structure. For ions in the ~1800 m/z range, the number of Al atoms is about 30 (with more Me and less O than this value). This many atoms provides an complicated structural challenge, as the connectivity possibilities are enormous. However, the problem is moderated by first ignoring the AlMe₃ (which are likely to be coordinated by O on the surface of the oligomer) and redecorating after a reasonable structure is established. We are using the bonding requirements of the Al and O (which form the framework of the oligomer) and symmetry as guiding principles in trying to establish reasonable structural models for these oligomers. We are also on the lookout for aluminum sites that may be Lewis acidic, because one is required for the demethylation step that generates the active catalytic species in olefin polymerization.

One postdoc (started 1 Feb 2010) and one MSc student (started 1 May 2010) have worked on the project to date (as well as small contributions in 2009 from a finishing PhD student and a part-time undergraduate).