Validating the Optical Properties of Machine Learning predicted Cu (I) Dipyrrin & Indole Complexes

Photodynamic therapy (PDT) shows promise in cancer treatment and diagnosis. Its mechanism involves the utilization of light to activate a photosensitizing agent, inducing malignant cell death. Advancing PDT outcomes require developing PSs with near-infrared absorption for deep tissue penetration. The most investigated transition metal complexes (TMCs) are made from precious metals with very low natural abundance in Earth’s crust. Copper is our choice of metal, and this is based on their wide availability and has not been reported toxic or carcinogenic if their accumulation in body is not abnormally high. To guide systematic design of improved PSs, we employ time-dependent density functional theory (TD-DFT) calculations on Cu (I) dipyrrin & indole complexes, aiming to elucidate structure-property relationships critical for enhanced PDT performance. Calculations show that these complexes exhibit a low-energy absorption peak appearing in the red-to-near infrared (NIR) regions in the range 600-1400 nm, this energy range is tunable by substituting electron withdrawing (CN, NO2) and electron donating groups (OCH3, NPh2), as well as changing the π-conjugation via side linking groups (H, -CH2,-CH2CH2-, -CH=CH-). The substituent groups change the charge transfer character of the low-energy excitons, while the conjugated connectors increase the degree of delocalization of the excitons. These compounds have been employed to analyze and generate virtual libraries of hypothetical near infrared (NIR) TMC systems for candidate screening and rational design using ML/cheminformatics methods. This would serve as a guideline for novel synthetic strategies of TMCs with desired NIR properties.