Predicting Supramolecular Stability: A Validated Framework Combining Orthogonal Topological and Electrostatic Descriptors

The design of functional liquids like Deep Eutectic Solvents (DES) is hampered by an incomplete understanding of the link between molecular components and the physical properties of the resulting mixture. Traditional metrics like binding energy (ΔE)fail to explain why some strongly bound pairs form low-viscosity liquids while others form glassy solids. This work introduces a holistic, multi-scale computational framework, built upon a validated quantum chemical method, to deconstruct supramolecular stability into two distinct, orthogonal physical phenomena: local bonding quality and global electrostatic neutralization. We develop two novel descriptors: the topological Non-Covalent Interaction Score (SNCI), which quantifies localized interaction strength, and the electrostatic COSMO Dimer Interaction Score (SCDI), which measures the homogeneity of the complex's surface potential. By applying this framework to nearly 3,000 DES complexes, we construct a "Stability Phase Space" that classifies pairs into distinct archetypes, such as rigidly bound "Local Champions" and flexibly packed "Global Stabilisers." This classification provides a new, physically-grounded rationale for designing complex liquids and offers a powerful, scalable methodology for screening candidates.