(366d) An Independent Assessment of the Anrtl Models.

Over the past 5 years Chen and co-workers have presented several versions of the adsorption NRTL (or aNRTL) model for both unary and multicomponent equilibrium adsorption isotherms, claiming novelty, thermodynamic consistency and improvements over traditional approaches.

Here we present an analysis of the aNRTL models, which are better described as the antisymmetric NRTL activity coefficient models. This property gives the models implicit asymmetry which results in negative deviations from ideality. For pure component isotherms the model is not new, it is in fact a variant of the Vacancy Solution Theory (VST). It will be shown that the structure of the model leads to heats of adsorption that are either decreasing with loading or nearly constants, thus excluding a large number of simple systems where nonpolar adsorbates that have adsorbate-adsorbate interactions have heats that increase with loading.

The multicomponent versions have a requirement of constant saturation capacities which are not imposed in any of the publications that have used the thermodynamic Langmuir model, thus making the model thermodynamically inconsistent, contrary to what is claimed. The apparent lack of dependence of the activity coefficients on the reduced grand potential can be reconciled if one understands which reference states are actually used, i.e. that this is a VST approach. This also explains why when incorrectly adding a reduced grand potential dependency limited improvements are observed.

The analysis is then extended to the generalized Langmuir model, a variant that includes the possibility for molecules to have different sizes. Worryingly, the expressions for the activity coefficients derived for this model appear to be incorrect, failing simple consistency checks which will be outlined in detail. In addition to this, it is interesting to note that the regressed model crosses the IAST predictions, but this is simply a consequence of the fact that the original data are thermodynamically consistent. Applying the integral consistency check to the model does reveal the problem.

We conclude with a set of recommendations that should be used when presenting new activity coefficient models for adsorption equilibria.