(366c) Thermodynamic Consistency Begins with Pure Component Isotherms.

Multicomponent equilibrium adsorption isotherms are one of the essential input variables required for the design of adsorption-based separation processes. However, due to the limitations of experimental methods and the vast possible combinations of multicomponent mixtures, it is impractical to determine them experimentally. As a result, adsorption process simulations rely on multicomponent predictions using analytical isotherms which are parametrized using pure component data. The importance of thermodynamic consistency in pure component model regression is paramount for the purpose of generating accurate multicomponent predictions.

Here we focus initially on multisite Langmuir isotherms. The task of fitting pure component isotherm models to equilibrium data is not trivial given the presence of local minima (and multiple global minima for multi-site models). It requires strong initial parameter guesses and a good optimizer. The requirement for thermodynamic consistency across adsorbate molecules further complicates the task and imposes the need for a constant saturation capacity for each site. On the other hand, imposing thermodynamic consistency results in a model that predicts an heterogeneous ideal adsorbed solution without the need to solve the HIAST equations.

In this contribution, we present a multi-step approach to fit equilibrium isotherms in a thermodynamically consistent manner using (the Virial or Myers isotherms). These are powerful tools for this purpose as they can be regressed by direct calculation, without need for optimizers or starting parameters. Thermodynamic quantities such as the enthalpy of adsorption, the Henry’s Law constant and the saturation capacity can then be determined from the fits and imposed to reduce the number of parameters to be determined and help yield a physically consistent fit of the thermodynamic adsorption isotherm model in an automated manner.

A final discussion of the system CO₂/N₂ on 13X zeolite will be used to highlight the importance of understanding the importance of selecting the correct ranges of temperature and pressure in the pure component data in order to obtain reliable parameters of the multisite Langmuir isotherm.