• • Solubility of β-HMX in DMSO–alcohol mixtures increases with temperature and DMSO mole fraction; solvent composition exerts a stronger influence than temperature, with a maximum relative solubility change of 6.3% due to composition fluctuation (equivalent to 1.4–2.7 K temperature shift). This underscores the need for precise solvent ratio control in industrial crystallization to avoid batch-to-batch variability.
• • The Apelblat model achieved the highest correlation accuracy with ARD < 5% and RMSD < 0.11%, outperforming Jouyban–Acree and NRTL models. This model can be reliably used for solubility prediction in process design and scale-up.
• • Thermodynamic analysis (NRTL-based) revealed that dissolution of β-HMX in all three solvent systems is endothermic (positive enthalpy), entropy-driven (positive entropy), and spontaneous (negative Gibbs free energy). This indicates that higher temperatures and higher DMSO content enhance solubility, guiding the selection of operating conditions for cooling or anti-solvent crystallization.
• • PXRD and DSC confirmed that β-HMX remains in its β-polymorph across all tested conditions, ensuring that solubility data are not confounded by polymorphic transitions. This is critical for reproducible crystallization and product quality control.