Romano, Eugenio (2025) Reaction modelling in computational chemistry: from DFT to high-accuracy methodology for improving catalysis knowledge. [Tesi di dottorato]

Documento PDF romano_eugenio_37.pdf Download (16MB)

Abstract

This thesis has pursued a unified computational view of polymerization catalysis that connects α-olefin and cyclic esters through the common concepts of stereoselectivity, regioselectivity and mechanistic aspects, with an eye to sustainable materials design. The objectives have been to investigate the mechanisms and the stereoelectronic effects that govern mechanism itself and microstructure of the final polymers, to build a benchmarked quantum-chemical protocol that reproduces experimental trends, and to derive design rules that transfer across ligand families and monomer classes. The methodology adopted combines density functional theory (DFT) with the Activation Strain Model and Natural Energy Decomposition (ASM-NEDA) to separate strain and interaction along the reaction coordinate. Steric maps based on buried volume quantify ligand congestion around the metal centre. For α-olefins, through DFT analysis the generally accepted Cossee mechanism has been revisited, predicting for low-symmetry systems a chain-stationary mechanism, arising from two diastereotopic sites. The origin of this mechanism has been studied and seems to be system dependent across bridged metallocenes, salalen and pyridylamido complexes. Site epimerization and diasterotopic sites are central. Steric maps alone do not account for the calculated energy gap corresponding to the experimentally observed selectivity, which motivates for an explicit decomposition of strain and interaction terms, making the former the determinant factor. Stereoselectivity is rationalized within the same picture, integrating two stereoerror pathways, one arising from a misoriented chain with anti insertion and one in which the preferred orientation of the growing chain is preserved while the monomer inserts syn. ASM–NEDA shows stereocontrol dominated by strain with the catalyst fragment as the principal discriminator, while the interaction term adds a secondary stabilization. On the same coordinates, resioselectivity is analysed, with primary 1,2 insertion favoured by interaction stabilization and strain factors stabilizing the secondary 2,1 insertion. The balance between steric and interaction terms sets the preferred regiochemistry. A dedicated benchmark links theory and experiment on propene stereocontrol across representative metallocene and post-metallocene families. Experimental pentad distributions are converted to free energy gaps and directly compared with computed barriers. Performance depends strongly on the treatment of dispersion in the optimization step and on solvent model in single point calculations. The same framework extends to cyclic esters. Achiral aluminium complexes control tacticity in rac-lactide through the monomer that determines the wrapping mode, creating local chirality at the active site. This aspect has been analysis through DFT, ASM-NEDA and buried volume analysis. A recently synthesised pyridylamido iron catalyst polymerizes δ and ε "challenging" lactones with high activity and narrow dispersity. Computations reproduce relative rates through a precise analysis of the mechanisms and support a picture in which different species, arising from the neutral precursor, can initiate the polymerization. Overall, the study delivers a view that bridges the worlds of polyolefins and polyesters, comparing experiments and theory, suggesting possible guidelines that link the main features of catalysts, including ligand nature and symmetry, to selectivity and activity, also in support of a circular polymer strategy.

Downloads

Actions (login required)

Modifica documento