Revitalized interest in biorenewable materials has revealed some accompanying challenges. For example, many compounds of interest, such as alcohols, are polar and readily self-associate, causing them to behave in a non-ideal manner. Equations of state (EOSs) such as the statistical associating fluid theory (SAFT), cubic plus association (CPA), and Elliot-Suresh-Donohue (ESD) are attractive options for modeling because they explicitly account for hydrogen bonding. However, these EOSs are typically parameterized by fitting macroscopic pressure-volume-temperature data, a practice that ignores molecular measurements of the bonding. Advancing the predictive power of thermodynamic models for polar systems requires molecular-level awareness, which can be provided by spectroscopy. This work implements variable-temperature infrared spectroscopy guided by insight from computational quantum mechanics to quantify the extent of hydrogen bonding in alcohol + cyclohexane systems based on the alcohol's hydroxyl stretching vibration. A new scaling technique is developed that provides for the first time a temperature-independent integrated area for the hydroxyl stretching region. For further validation of the new scaling method, the scaled infrared spectra are correlated to the nuclear magnetic resonance spectra for 1-butanol + cyclohexane and 2-propanol + cyclohexane using quantum calculations with minor empirical adjustments. The infrared measurements are used to parameterize two association constants for each binary system, which are implemented in a new activity coefficient model based on the resummed form of Wertheim's perturbation theory (RTPT). The widely used implementation of one association parameter for each binary (TPT-1) in PC-SAFT, CPA, and ESD is shown to be inadequate for fitting the spectroscopic data. The RTPT model succeeds in recovering the hydroxyl bond type distributions from the infrared measurements. When the association constants from spectroscopy are applied to the modeling of phase equilibria, association is demonstrated to be the dominant contribution to solution non-ideality. When combined with combinatorial and residual models, RTPT provides an improved representation of experimental phase equilibria and excess enthalpies when compared to the TPT-1 model.
Read
