(437b) Kinetic Modeling of Single-Site Olefin Polymerization with Multi-Response Data: Even Models with Many Parameters Cannot Fit An Elephant

Single-site catalysts are revolutionizing polyolefin production because they enable more precise control of the polymer's molecular architecture, which in turn controls the physical properties of the polymer. The ability to design single-site catalysts requires fundamental understanding of the polymerization, including detailed knowledge of the governing kinetic mechanisms and values for all of the rate constants. This task is difficult because for each of the individual steps in the polymerization (i.e. activation, initiation, propagation, and chain transfer) there are a number of mechanistic possibilities, resulting in potentially thousands of different polymerization models. Hence, it would seem that any number of proposed kinetic models would be capable of describing the polymerization, because of the large number of kinetic parameters associated with each potential model. On the contrary, we have found that for a rich set of multi-response data (i.e. evolution of monomer concentration, molecular weight distribution, different types of terminal groups via NMR, etc) it is actually very difficult to fit all the data with a mechanistically based kinetic model ? one cannot fit the proverbial elephant no matter how many parameters if the proposed mechanism, and hence associated kinetic equations, is in error. Thus, quantitative kinetic modeling provides an essential tool for determining detailed mechanisms.

The power of quantitative kinetic modeling in combination with rich multi-response experimental data will be demonstrated here for polymerization of 1-hexene with the [rac-(C₂H₄(1-Ind)₂)ZrMe][MeB(C₆F₅)₃] single-site catalyst system. Experimental batch polymerization data include (i) monomer concentration versus time profiles via ¹H-NMR, (ii) time evolution of the molecular weight distribution, (iii) terminal vinylene and vinyledene concentrations via ¹H-NMR, (iv) structure of the polymer attached to catalysts obtained by quenching with deutero-methanol with subsequent ²H-NMR analysis and (v) electronic structure of the metal-counterion complex from UV-VIS. This set of rich multi-response data, in combination with quantitative kinetic analysis, indicates that the mechanisms proposed in the literature for this catalyst are inadequate to describe all the observations. A new mechanism with a dormant site that can reactivate is required to describe the data. This communication will demonstrate how this approach can be used to discriminate between possible mechanisms, where the researcher can be confident in a kinetic model once it is able to fit all the multi-response data.