519541 The Role of Competitive Sorption in CO2/CH4 Separation with Glassy Membranes

Tuesday, June 12, 2018: 4:00 PM
Grand Ballroom Salon D (Hilton Lexington)
Eleonora Ricci, DICAM and INSTM, Bologna, Italy, Francesco Maria Benedetti, Department of Civil, Chemical, Environmental and Materials Engineering (DICAM), University of Bologna, Bologna, Italy and Maria Grazia De Angelis, Department of Civil, Chemical, Environmental and Materials Engineering, University of Bologna, Bologna, Italy

The mixed gas transport of CO2/CH4 mixtures in glassy polymers suitable for natural gas purification like Matrimid, Cellulose Acetate, PIMs and Thermally Rearranged (TR) polymers deviates significantly from their pure gas behaviour, leading to large uncertainty in the evaluation of the real membrane performance in different operative conditions. Multicomponent phenomena that affect the permeation and separation of this mixture are competitive sorption, plasticization induced by high CO2 pressure, swelling. CO2 is smaller than CH4, and therefore the diffusivity selectivity for such mixture is higher than unity, but the size difference is not high, and most of those polymers base their separation performance on a significant value of solubility selectivity.

While the diffusivity selectivity is mostly hard to predict a priori on the basis of pure component experimental data, models like the Dual Mode and NET-GP for glassy polymers have expressions for the estimation of mixed gas solubility and solubility-selectivity based on pure component sorption data.

Over the years, we have collected a number of multicomponent solubility data for the CO2/CH4 mixture in many ultraglassy polymers, identifying relevant deviations between real and ideal sorption behaviour, and generalized trends that are suitable for membrane design.
In this work we compare the trends of real solubility-selectivity with those of the available real permeability- selectivity data for many glassy polymers, trying to identify the origin and nature of gas-gas-polymer interaction and establish generalized correlations.


Extended Abstract: File Not Uploaded