Viscosity modeling of propyl butanoate and 2-alkanol mixtures using modified Cohen-Turnbull and UNIFAC-VISCO...

Densities and dynamic viscosities of binary mixtures containing propyl butanoate (PB) with 2-propanol, 2-butanol, 2-pentanol, 2-hexanol, and 2-heptanol were measured at temperatures ranging from 293.15 K to 323.15 K and ambient pressure. Two distinct modeling approaches were developed and rigorously evaluated against the experimental data: the Modified Extended Cohen-Turnbull Model (M-ECTM) and the Excess Free Volume Corrected UNIFAC-VISCO (EFVC-UNIFAC) model. While the standard UNIFAC-VISCO group-contribution method exhibited substantial systematic deviations (global AARD = 8.0%, bias = - 6.8%) due to its inability to capture hydrogen-bonding disruption and volume expansion effects, both proposed models achieved significant improvements. The M-ECTM, incorporating composition-dependent activation barriers via Redlich-Kister expansion, yielded the highest overall accuracy with a global AARD of 2.9% and MaxAD of 5.8%, utilizing three adjustable parameters per binary system. Conversely, the EFVC-UNIFAC approach, which introduces a single corrective term proportional to the experimental excess molar volume (V^E), achieved comparable precision (global AARD = 3.1%, MaxAD = 7.2%) with only one binary parameter (ξ₁₂). This parameter exhibited an increasing trend with increasing alcohol chain length (0.048 for 2-propanol to 0.084 for 2-hepatanol). Both models effectively eliminated systematic bias (|bias| < 0.5%), with the M-ECTM performing marginally better for longer-chain alcohols and the EFVC-UNIFAC demonstrating superior statistical efficiency. These results establish that the M-ECTM is optimal for maximum precision, whereas the EFVC-UNIFAC offers an excellent compromise between accuracy and parsimony for engineering applications involving ester-alcohol mixtures.