Spectroscopic Fingerprints of MgCl₂/TiCl₄ Nanoclusters Determined by Machine Learning and DFT

Understanding the structure and properties of MgCl₂/TiCl₄ nanoclusters is a key to uncovering the origin of Ziegler-Natta catalysis. In particular, vibrational spectroscopy can sensitively probe the morphology and active species of MgCl₂/TiCl₄. Here, we determined vibrational spectroscopic fingerprints of 50MgCl₂ and 50MgCl₂/3TiCl₄ which were obtained by nonempirical structure determination based on an evolutionary algorithm and DFT. The adsorption of CO, TiCl₄, and Ti₂Cl₈ dimers was also modeled on each of the coordinatively unsaturated Mg²⁺ sites available for binding including so-called defect sites, which are likely present at the surface of activated MgCl₂ nanocrystals and plausible sites for strong TiCl₄ species adsorption. The outcomes of thermodynamical and vibrational analysis were compared to results on ideal surfaces of MgCl₂. Vibrational analysis (IR and Raman) on plausible models of TiCl₄/ MgCl₂ nanoclusters revealed that IR response is useful for distinguishing between the different ways of binding of TiCl₄ on different sites of adsorption, whereas Raman response provides a clear fingerprint of supported TiCl₄ species.