A unified solvatochromic-computational-machine-learning approach for solvent property prediction based on iodine color response.

Solvent effects govern the majority of chemical transformations, yet experimentally accessible and universally applicable tools for predicting solvating ability remain limited. Here, we introduce iSolv, a simple but information-rich solvatochromic descriptor derived from the color of diluted iodine solutions in organic solvents. Using a combined experimental-computational approach, we demonstrate that iSolv directly reflects the nature and strength of intermolecular interactions, including dispersion interaction, halogen bonding, donor-acceptor interactions and hydrogen bonding. We establish clear correlations between iSolv and key electronic, physicochemical, and empirical solvent parameters, such as HOMO energy, donor number, log P, Kamlet-Taft parameters, and reaction rates in diverse organic transformations. Machine-learning modelling confirms that iSolv enables predictive assessment of solvent behavior using easily accessible input parameters. Unlike existing multi-parameter solvent scales, iSolv offers a low-cost, rapid, and visually interpretable method applicable across chemical research, education, and chemical technology. The practical and computational relevance of iSolv positions it as a versatile tool for solvent selection, reaction optimization, and broader chemical informatics. This work introduces a convenient indicator-based strategy for assessing solvating ability and lays the foundation for integrating visual color response into the digital chemistry framework.