A new activity coefficient model for describing adsorbed phase nonidealities

Abstract

The following study introduces a new adsorbed phase activity coefficient model, the SPLINT model, for describing nonideal multicomponent systems. The derivation is presented and simplification to the thermodynamically robust ABC model in the low‑coverage limit is shown. Both the SPLINT and ABC models are fit to 25 binary systems. Thermodynamic consistency of the data is established through multiple consistency tests including the intersection rule, closed‑loop reduced spreading pressure integration, the Gibbs‑Duhem integral equality, and alignment of pure and binary data. Regressed parameters are used in Real Adsorbed Solution Theory to predict both binary and ternary adsorption. The ABC and SPLINT models accurately describe systems with considerable nonideality when the activity coefficients are symmetric; however, the SPLINT model is required for accurate description of systems with complex, asymmetric activity coefficient behavior. The new model includes one more fitting parameter than the ABC model for a single binary system, but less across multiple data sets and higher‑order systems due to the theoretical basis. The importance of capturing activity coefficient asymmetry is demonstrated with a highly nonideal system across various pressures and temperatures measured by the present authors using the Integral Mass Balance method with HFC‑125 (pentafluoroethane), HFC‑32 (difluoromethane), and zeolite H‑ZSM‑5. The theoretical basis of the SPLINT model is further enforced by presenting two methods for predicting activity coefficients with the SPLINT model: correlating the fitting parameters individually and simultaneously regressing pure isosteric heat data to extract SPLINT fitting parameters. The study concludes with a discussion of pure and binary HFC‑125/HFC‑32 data with zeolite H‑ZSM‑5. A possible binary adsorption mechanism is proposed based on isosteric heat and excess enthalpy predictions and observed system irreversibility.