Development of Data-assisted Photo-Organocatalytic Transformations

Photocatalysis is now a well-established method for developing sustainable protocols in synthetic organic chemistry. Organophotoredox catalysis, in particular, is gaining significant interest due the use of inexpensive, easy-to-make, metal-free organic dyes with tunable photophysical properties. Structural modification of these organic molecules is crucial to fine-tune their redox features for desired photochemical transformations. However, predicting the final excited-state properties of functionalized dyes remains challenging, especially when compatibility with the reaction system is required. This has led to a resource-intensive trial-and-error approach to identify the optimal photocatalyst structure for specific transformations, limiting its widespread application. A promising solution is data science, which can streamline the relationship between the chemical structure of dyes and their function, facilitating their application. The computer-assisted development of innovative systems in organophotoredox catalysis can offer novel possibilities for large-scale applications, reducing time, reagent consumption, and costs compared to traditional methods. This proposal aims to develop new data-assisted protocols for reaction discovery and photocatalyst design. During the outgoing phase at the University of Utah, under Prof. Sigman's supervision, machine learning algorithms will be used to correlate photocatalysts' structures with their functions. Classification tools will explore reactivity cliffs, and dimensionality reduction techniques will map the chemical space to visualize reactivity patterns with fewer experiments. In the subsequent return phase at the University of Padova, under Prof. Dell'Amico's supervision, these photochemical methodologies will be adapted for flow processes to produce synthetically relevant compounds on a multi-gram scale. The data science knowledge acquired will guide this transition, enabling a more efficient implementation.