interactions need to be included. Hence, many variants of SAFT have been developed
depending upon the way attractive interactions∕dispersions are included in the model.
In the original SAFT publication, Chapman et. al. [46] included a generalized
van der Waals mean field term to account for the dispersive forces. Such a simple
dispersion term is, however, not suitable for modeling the properties of real polymeric
fluids. Huang and Radosz [63, 64] developed the first widely applied version of SAFT,
known as HR-SAFT. The dispersion term is based on a power series fitted to accurate
PvT, internal energy, and second viral coefficient data for argon, by Chen and Kre-
glewski [65]. Huang and Radosz obtained pure component parameters for over IOO
non-associating and associating components. The fitted parameters are well-behaved,
and for a homologous series such as n-alkanes, they follow a trend. This allows users
to estimate the pure component parameters for larger molecules. This is the main
reason for HR-SAFT being widely used in the 1990’s, and it is still applied today.
Another approach that has been followed is to use Lennard-Jones (LJ), Square
Well (SW) or Yukawa potential to define the attractive (spherical) reference fluid and
bond them to form chain fluids. Chapman [66] proposed a EOS for LJ chain fluids (LJ-
SAFT) which has been extended and applied to real fluids by Bias and Vega [67, 68]
and Kraska and Gubbins [69, 70]. The EOS developed by Bias and Vega is labelled
as soft-SAFT and has subsequently been applied to mixtures of hydrocarbons [71],
perfluoroalkanes [72], and recently to polymers [73]. Similar ideas were employed by
Banaszak et. al. [74] to propose an EOS for SW chain fluids.
33