Comprehensive Modeling of Acetone Clusters: QTAIM Analysis and QCE Study

Abstract

ABSTRACT In molecular research, comprehending the microscopic source of the macroscopic characteristics of polar aprotic solvents continues to be a significant difficulty. In order to bridge the gap between cluster‐scale interactions and liquid acetone properties, we present a thorough quantum‐chemical and statistical modeling of neutral acetone clusters in this work. The ABCluster algorithm was used to thoroughly explore the potential energy surface. High‐level density functional theory calculations at the MN12SX‐D3/def2‐TZVP level were then performed, benchmarked against DLPNO‐CCSD(T)/CBS reference energies. A thorough Quantum Theory of Atoms in Molecules (QTAIM) analysis of the nature and hierarchy of intermolecular interactions revealed a cooperative network dominated by dipole‐dipole O⋯C and O⋯O interactions, supplemented by numerous weak C‐H⋯O, H⋯C, and H⋯H dispersive contacts. The application of the QCE theory predicts a distribution dominated by trimers at low temperatures (T< 200 K), leading to a predominance of monomers above 260–280 K, reflecting the subtle equilibrium between electrostatic stabilization and entropic effects. The model reproduces experimental thermodynamic properties, such as the thermal capacity (Cp) between 200 and 375 K and infrared spectra at 300 K, with the calculated band of elongation C=O (1710 cm –1 ) being just 5 cm –1 from the experimental value (1715 cm –1 ). The thermodynamic properties and infrared spectrum of liquid acetone predicted by QCE show excellent agreement with experimental data, thus validating the integrated DFT‐QTAIM‐QCE approach. This work provides the first complete QCE characterization of pure liquid acetone, demonstrating that its macroscopic properties emerge from a dynamic equilibrium of small, weakly‐bound clusters rather than extended hydrogen‐bonded networks, and establishes a validated computational framework for predicting liquid‐phase properties from ab initio cluster data.