American Chemical Society
Browse

Statistical Analysis of Catalytic Performance in Ethylene/Methyl Acrylate Copolymerization Using Palladium/Phosphine-Sulfonate Catalysts

Download (4.97 MB)
journal contribution
posted on 2022-03-29, 21:04 authored by Shumpei Akita, Jing-Yao Guo, Falk W. Seidel, Matthew S. Sigman, Kyoko Nozaki
For various types of palladium complexes bearing phosphine-sulfonate (PS) ligands used in the coordination–insertion copolymerization of olefins with polar monomers, characteristic features of the ligands, such as electronic and steric properties, have been discussed to describe their catalytic performance. Aiming at further analysis of the literature data, here we report the development of a statistical method for how the ligand impacts the performance of a Pd-catalyzed copolymerization of ethylene and methyl acrylate (MA). During our investigation, ligand features important for the resultant molecular weight of the obtained polymers were identified. Consistent with previously suggested important parameters, the electron density on the palladium center and maximum width of the substituents on the phosphorus atom (B5) were found to be significant for catalyst performance. We also found that additional features impact reaction outputs. As an example, the lower occupancy of the palladium dz2 orbital results in an increase of molecular weight and catalyst activity in both ethylene homopolymerization and ethylene/methyl acrylate copolymerization. Furthermore, it was predicted that a larger bite angle of the ligand increased the activity of ethylene/methyl acrylate copolymerization without impacting the molecular weight. On the basis of these machine learning predictions, three thiophene derived PS-type catalysts were synthesized and tested for MA/ethylene copolymerization. Unexpectedly, rather than the one predicted to enhance catalytic performance, a synthetic intermediate to this ligand exhibited higher activity albeit with the expense of molecular weight and MA incorporation. The inconsistency between the prediction and the experimental result is likely a result of insufficient training data for the catalyst with a different linker moiety. However, the unexpected finding that chlorination of the ligand backbone increases the overall catalyst performance will inspire an avenue for PS catalyst development.

History