A novel mixing rule that bridges the Statistical Associating
Fluid
Theory (SAFT)-type equations of state and activity coefficient models
is proposed. Applying this mixing rule to the PC-SAFT equation of
state is the focus of this article. In comparison to the original
PC-SAFT equation, and even the underlying activity coefficient model,
this mixing rule provides high-accuracy predictions of equilibrium
properties for a range of mixtures, being able to predict phenomena
that neither PC-SAFT or the underlying activity coefficient model
would be able to predict on their own. A few limitations are identified
including the case of cross-associating mixtures, due to the difficulty
of separating associative interactions from dispersive interactions
in activity coefficient models. Nevertheless, it is also shown that
the new mixing rule is able to predict bulk properties very accurately,
including volumetric properties which activity coefficient models
alone are not able to predict. Given that one is able to use predictive
activity coefficient models within this mixing rule, such as UNIFAC
and COSMO-SAC, this new mixing rule opens the doors for the development
of fully predictive SAFT equations of state for mixture systems.