A global molar mass capability model for early transition metal olefin polymerization catalysts

In the framework of various projects funded by the Dutch Polymer Institute we have tested multiple catalyst classes (metallocenes, Salan catalysts, [OOOO] catalysts, Mitsui-type catalysts) in propene polymerization catalysis to isotactic polypropylene (iPP) under judiciously identical conditions employing our high-throughput experimentation (HTE) screening techniques.1-3 The combined database (Tp = 100°C, >100 catalyst) spans catalysts with molar mass capabilities ranging from oligomers to ultrahigh-molecular weight iPP. Importantly, for most catalyst classes examples for minimal, average, and ultrahigh molar mass capability are included through variation of the substituent patterns. Similar to our previous approach which focused solely on the performance of zirconocenes,1-2 the combined HTE database was used to develop a regression model using a minimal set of intuitive chemical descriptors that were developed for stereoselectivity, regioselectivity, and comonomer affinity. The out-of-the-box performance of this model is demonstrated as no new descriptors are needed. Several important observations can be made: (a) electronic effects appear to play at best a minor role, (b) in difference to other recent VBur approaches,4 steric bulk in different regions affects performance differently, i.e., a single buried volume descriptor centered on the metal cannot predict the performance accurately; a global model requires deconvolution of regions of the active pocket. This research forms part of the research programme of DPI, project #835.