Development of Photochemical Strategies for the Selective Functionalization of pi-Extended Systems Using Bayesian Learning

• Grant number: 25/12870-7
• Support Opportunities: Scholarships in Brazil - Doctorate (Direct)
• Start date: November 01, 2025
• End date: July 31, 2030
• Field of knowledge: Physical Sciences and Mathematics - Chemistry - Organic Chemistry
• Principal Investigator: Marco Antonio Barbosa Ferreira
• Grantee: Camila Pelizzer de Oliveira
• Host Institution: Centro de Ciências Exatas e de Tecnologia (CCET). Universidade Federal de São Carlos (UFSCAR). São Carlos , SP, Brazil
• Associated research grant: 21/12394-0 - Sustainable synthetic methods employing catalysis, benign solvents, safer reagents, and bio-renewable feedstock, AP.TEM

Abstract

Methodologies combining palladium complexes and visible-light activation have emerged as versatile and selective alternatives for the functionalization of a wide range of molecular systems, particularly by enabling the generation of radical intermediates under mild conditions. Preliminary studies conducted by our group have demonstrated the feasibility of ¿-selective couplings of ¿-extended silicon enolates derived from aromatic ketones, employing alkyl halides as radical precursors and light-activated Pd catalysts. These transformations, unprecedented in the literature, proceeded with high regio- and stereoselectivity but still face limitations in terms of generality and reproducibility. The central goal of this project is to broaden the synthetic scope of these reactions by integrating Bayesian Optimization algorithms to iteratively and data-drivenly guide the selection of more efficient reaction conditions. We also aim to apply the methodology to a more diverse set of substrates-including enolates derived from esters and aliphatic ketones-and to explore new radical precursors such as alcohols and carboxylic acids, with a strong focus on sustainability. In parallel, mechanistic investigations will be carried out using multivariate linear regression (MLR) models based on molecular physicochemical descriptors obtained both experimentally and computationally. This approach is expected to establish structure-reactivity relationships, rationalize trends observed during optimization campaigns, and contribute to the development of predictive principles in Pd-mediated photoredox catalysis.