Navigating the Solubility Landscape of APIs in Deep Eutectic Solvents: A Data-Driven Thermodynamic Taxonomy of...

Deep eutectic solvents (DESs) have emerged as powerful media for enhancing the solubility of poorly water-soluble active pharmaceutical ingredients (APIs). However, their rational design remains challenging due to the complex interplay of intermolecular interactions and non-ideal thermodynamic behavior. This study develops a comprehensive, data-driven taxonomy of solute-solvent systems by integrating COSMO-RS-derived descriptors with principal component analysis (PCA) and unsupervised clustering. This approach establishes a constrained, evidence-based decision framework, which is more appropriate for complex physicochemical systems like DESs than traditional empirical rules. The analysis successfully reduces the multidimensional descriptor space to five physically interpretable axes: solvation driving force, API thermodynamic stability, solvent interaction profile, hydrogen-bond network strength, and hydration effects. Two primary solubilization mechanisms are identified: interaction-driven solvation, characterized by high API-DES affinity, and destabilization-driven solvation. Furthermore, comparison of dry and water-containing systems reveals that water acts as a thermodynamic structuring agent, fundamentally reducing system dimensionality and promoting the emergence of more distinct solvation regimes. Validated through the projection of benzocaine and lidocaine, this framework enables a transition from trial-and-error screening to mechanism-guided formulation design, providing a robust roadmap for navigating the complex solubility landscape of pharmaceutical DESs.