The SAFT family of equations comprises some of the very few equations of state which explicitly account for the hydrogen-bonding interactions that exist in associating fluids such as water and alcohols. Two such equations are the Cubic-Plus-Association (CPA) and the recently developed Perturbed-Chain-SAFT (PC-SAFT) equations of state. Both these equations of state require 5 pure-component parameters to be specified for each associating fluid. Three of these are physical parameters (also required for non-associating fluids) whereas the remaining two parameters characterize the hydrogen bonding sites on the molecule. These five parameters are typically obtained by simultaneous fitting to experimental saturated liquid density and vapour pressure data. When so many parameters are required, the physical meaning of the parameters may be lost in the estimation procedure. It would thus be useful if further experimental data, specific to association fluids, could be used, when obtaining parameters for these fluids. Equations of state based on the Wertheim formalism (such as the SAFT variants) naturally calculate the degree of hydrogen bonding that exists in a pure fluid or mixture at given conditions. But IR and Raman spectroscopy can give us just such information experimentally.

In this work we investigate the ability of both CPA and PC-SAFT to predict the degree of hydrogen-bonding in a number of pure associating fluids (water and 1-alkanols) over an extended temperature range. The effect of parameter estimation on these predictions was tested and found to be marked. It is thus clear that spectroscopic data can be used to obtain physically more meaningful parameters without comprising accuracy in predicting other, more useful, physical properties. CPA and simplified PC-SAFT (recently developed in our laboratory) were then used to predict the degree of hydrogen-bonding in binary mixtures consisting of an associating and a non-associating fluid. In general, both equations of state gave reasonable agreement with experimental spectroscopic data using parameters obtained only from density and vapour pressure data. The agreement could be improved by incorporating spectroscopic data into the parameter fit. Additional experimental data for binary systems have also been obtained using FTIR spectroscopy. Finally, in the literature there exist a number of inconsistencies arising from the confusion between non-bonded sites on a molecule and completely non-bonded molecules. These anomalies are pointed out and corrected.