Infinite Dilution Revisited: Filling Predictive Gaps in Physical Property Models

All popular process simulation programs contain thermodynamic packages that can contribute to development and design of separation systems involving distillation. Still, there is nothing completely automatic about selecting an appropriate physical property package containing the best parameters for the job at hand. If we choose poorly, the method may not yield the accuracy we require. Often there is no one package that will handle all of the components well.

This talk will illustrate this challenge for mixtures containing hydrogen-bonding components. We will demonstrate how straightfoward application of simple molecular reasoning can lead to better property predictions than possible using the standard physical property correlations.

Two applications of special interest will be addressed. The first is the trapping of trace polar components in distillation which leads to column hiccups. The second application involves design opportunities where more accurate prediction of temperature dependence opens the door to alternative separation techniques, such as pressure-swing distillation.