Ziegler-Natta Catalysts: Mechanistic Study via High Throughput Screening Methodologies

Heterogenous High-Yield Ziegler-Natta catalysts (HY-ZNCs) are the most widely used systems for the industrial production of isotactic polypropylene (iPP) based materials, including the homopolymer, †˜random' copolymers, and reactor blends of homopolymer and ethylene/propylene †˜rubber' (†˜impact' PP). They consist of a support (MgCl2), a transition metal precursor (e.g. TiCl4), an activator (e.g. an Al-trialkyl), and one or more electron donor modifiers (e.g. an ester, ether, alkoxysilane). Despite 40 years of intensive research (60, if one includes first-generation TiCl3-based catalysts), the complexity of these formulations, in which subtle changes in composition, preparation and/or application protocols often result into dramatic effects in performance, prevented a rational design, and thus far the industrial progress was dominated by empiricism. In this thesis we carried out an intensive High Throughput Experimentation (HTE) study of HY-ZNCs, with the general aim to improve the understanding and control of active site structure and behavior. In particular, two state-of-the-art HTE platforms (namely, a Freeslate Extended Core Module and a Freeslate PPR48 setup), both integrally contained in a glove-box environment and integrated with advanced analytical techniques (such as NMR with high-temperature cryoprobe, ICP-OES, Rapid GPC, etc), were used to thoroughly investigate representative catalyst systems. A careful analysis of the adsorption/desorption processes that occur on the catalytic surfaces under conditions representative of industrial use, the correlation of said phenomena with the polymerization behaviors, and a suitable integration of the experiments with state-of-the-art periodic DFT-D modeling, led us to formulate, if not yet a working white-box model of these systems, convincing and reasonably well-defined hypotheses on the structure of the active species, and their non-bonded interactions with various adsorbates, including organic electron donors, Al-alkyls, and †"possibly†" reaction products thereof. On the other hand, the wide and robust experimental HTE database was employed in parallel to successfully implement a Quantitative Structure/Activity Relationship (QSAR) model of HY-ZNC surface modification by means of alkoxysilane electron donors, featuring predictive ability for the first time ever. Although admittedly of black-box character and limited to one of the several industrial catalyst platforms, based on a MgCl2/TiCl4/diisobutyl-ortho-phthalate precatalyst, this result represents the first case of computer-oriented HY-ZNC surface fine-tuning, and opens the door to the fast identification of novel and useful catalysts and polymers.