Modelling competitive sorption and plasticization of glassy polymeric membranes used in biogas upgrading

A mathematical model of membrane performance is developed that incorporates fugacity-dependent permeabilities, competitive sorption, penetrant blocking and plasticization effects. The model also accounts for nonisothermal operation and includes real gas behavior and concentration polarization. Importantly, the model simultaneously considers plasticization caused by water vapor (H2O) and carbon dioxide (CO2). A simulation of biogas (composed of methane (CH4), CO2 and H2O) upgrading is performed using the new model and compared to models that use constant and pure gas permeability. Relative to these simplified models, the new model predicts differences up to 2% and 18% in CH4 recovery at low feed flowrates and the difference in CO2 removal can be as significant as 50%. Furthermore, simulations with and without water vapor in the feed give predictions that are 4.5%-34% different. The differences are attributed to the changes in fugacity-dependent permeabilities, particularly the sensitivity of these permeabilities to feed composition. An analysis indicates that the contributions of competitive sorption and penetrant blocking/plasticization to these differences is 19% and 32% in terms of CH4 recovery and CO2 removal, respectively. The remaining differences are due to real gas behavior, while concentration polarization has a negligible impact under the chosen conditions.

Automated summary

A mathematical model of membrane performance incorporating various factors such as fugacity-dependent permeabilities, competitive sorption, penetrant blocking, and plasticization effects was developed. The model considers real gas behavior, concentration polarization, and the influence of water vapor and carbon dioxide on plasticization. Simulation results showed significant differences in methane recovery and CO2 removal between the new model and simplified models, with the new model showing higher accuracy in predicting these values.