CO2/CH4 mixed-gas separation in PIM-1 at high pressures: Bridging atomistic simulations with process modeling

Marcel Balçık, S. Birgül Tantekin-Ersolmaz, Ingo Pinnau, M. Göktuğ Ahunbay

Research output: Contribution to journalArticlepeer-review

Abstract

Polymeric membranes with intrinsic microporosity have been at the center of attention for gas separation applications since the introduction of PIM-1. This study utilizes atomistic simulations to model and to understand the pure- and mixed-gas transport properties of PIM-1 for the CO2/CH4 gas pair. Monte Carlo and molecular dynamics methods were combined in the estimation of sorption and diffusion of CO2 and CH4 in PIM-1. Simulated sorption and permeability data compared very well with experimental reports. Mixed-gas adsorption simulations proved the existence of competitive adsorption, favoring CO2, hence resulting in an increase in solubility selectivities. However, in mixed-gas environment CH4 permeabilities increased significantly compared to pure gas conditions, overall decreasing perm-selectivities of the polymer. Plasticization of the polymer around 25 bar CO2 partial fugacity was apparent both in pure- and mixed-gas conditions. Simulations at different gas feed compositions proved the dependence of competitive sorption and CO2-induced swelling in partial feed gas fugacities. Simulation results were combined to obtain a macroscopic permeability model that relates the multicomponent permeability to the permeate pressure and composition. Accurate estimations of permeabilities by the model were achieved allowing future implementation of the model in process simulation tools.
Original languageEnglish (US)
Pages (from-to)119838
JournalJournal of Membrane Science
DOIs
StatePublished - Sep 2021

ASJC Scopus subject areas

  • Biochemistry
  • Filtration and Separation
  • Materials Science(all)
  • Physical and Theoretical Chemistry

Fingerprint

Dive into the research topics of 'CO2/CH4 mixed-gas separation in PIM-1 at high pressures: Bridging atomistic simulations with process modeling'. Together they form a unique fingerprint.

Cite this