Predictive Approach for the Solubility and Permeability of Binary Gas Mixtures in Glassy Polymers Based on an NETGP-NRHB Model

The NETGP-NRHB lattice-fluid model describes the sorption thermodynamics of penetrants in amorphous polymer-penetrant mixtures locked in an out-of-equilibrium “glassy” state. It accounts for the nonrandom distribution of mean-field contacts and voids and for the possible occurrence of specific interactions. In this contribution, we assess the suitability of the NETGP-NRHB model to interpret the sorption thermodynamics of binary penetrant mixtures in glassy polymers. Moreover, adopting the gradients of NETGP-NRHB penetrant chemical potentials as driving forces for their diffusive fluxes, a self-consistent framework is proposed to predict the permeability of light gas binary mixtures in glassy polymer membranes once all of the model parameters are obtained by nonlinear regression of data of the corresponding binary and pure component subsystems. This approach is validated against solubility data of CO2/CH4 and CO2/C2H4 mixtures in glassy amorphous poly(2,6-dimethyl-1,4-phenylene oxide) (PPO) and poly(methyl methacrylate) (PMMA), respectively, and permeability data of CO2/CH4 mixtures in three glassy amorphous polymer membranes (PAr, PSf, PH); all of the data are taken from the literature. In particular, solubility data are investigated up to a CH4 partial pressure of 15 atm at a fixed CO2 partial pressure equal to 5.1 atm for the system CO2/CH4/PPO and up to a CO2 partial pressure of 4 atm at a fixed C2H4 partial pressure equal to 2.09 atm for the system CO2/C2H4/PMMA. Permeability data are investigated at a fixed upstream molar concentration of the binary gas mixture equal to 0.5 by changing the total pressure up to 18 atm for the systems CO2/CH4/PAr and CO2/CH4/PSf and up to 14 atm for the system CO2/CH4/PH; the downstream total pressure is equal to 0 atm for all of the systems. All of the experimental sets of data are obtained at 35 °C. Solubility and permeability predictions for the ternary systems compare very well with all experimental literature data without additional parameters besides those required for the corresponding binary subsystems.