Adsorption of light gases and gas mixtures on zeolites and nanoporous carbons

Abstract

The ability to accurately predict mixture adsorption equilibrium is vital for future gas separation technologies. This research concentrates on the investigation and prediction of adsorption equilibrium of pure components and gas mixtures. The systems studied are of importance to the field of personal medical oxygen generation. Two commercial adsorbents are investigated: a LiLSX zeolite and a carbon molecular sieve. Adsorption equilibria are measured on the LiLSX zeolite for nitrogen and oxygen, both as pure gases and binary mixtures. The binary measurements consist of Henry’s law behavior with one component in excess as well as binary isotherms for a range of compositions. Binary Henry’s law behavior is modeled by adding virial excess mixture coefficients to the ideal adsorbed solution theory. The mixture coefficients are evaluated solely from the binary Henry’s law behavior, and the binary isotherms for a range of compositions are predicted accurately. Adsorption isotherms are measured on the carbon molecular sieve for oxygen, nitrogen, and argon at pressures up to 100 bar. The capacities of the gases are higher than expected, with the carbon molecular sieve having capacities higher than BPL activated carbon and capacities similar to a superactivated carbon. An approach to model adsorption of pure gases using classical density functional theory is developed. Fundamental measure theory is used as an accurate method to describe the hard sphere interaction potential. To address the intermolecular and intramolecular attractive potentials, a version of the statistical associating fluid theory is used. The new approach models adsorption of molecules of increasing complexity on flat walls, with predictions comparing well with simulation data in the literature. The new approach is extended by incorporating the 10-4-3 Steele potential to describe slit-shaped carbon pores. Nitrogen is first modeled inside the pores, where the density profiles are combined to produce a pore size distribution and describe the nitrogen isotherm. Pore density profiles are modeled for n-pentane and, using the pore size distribution obtained from nitrogen, a pentane isotherm is shown to agree well with experimental data.