Evaluation of oxygen separation from air by pressure/vacuum swing adsorption

Abstract

This study aimed to evaluate the production of high purity oxygen (90–95%) through experiments in a PSA/VSA unit and develop a mathematical model to describe the dynamic behavior of the process. Different operational parameters and the dead volume were investigated for their impact on process performance. The experiments used a laboratory-scale unit filled with beads of a commercial LiX zeolite to obtain breakthrough and PSA/VSA data for model validation. Equilibrium isotherms of pure oxygen and nitrogen were measured at 288, 298 and 313 K for the pressure range of 0 to 3 bar. Single and multicomponent breakthrough curves were obtained at 298 K. Synthetic air (grade 5.0 purity, excluding argon) with a composition of 20% (± 0.5%) O₂ and 80% (± 0.5%) N₂ was used in the PSA/VSA experiments. A novel approach was developed using the mathematical model designed to simulate PSA/VSA cycles to account for the dead volume effects commonly found in units of this type. The model was implemented and solved using gPROMS® software. The simulation data matched well with the experimental data, accurately representing histories of concentration, pressure, temperature, and purity variations during the process. The validated model revealed optimal operating conditions for a VSA unit: 7.5 s adsorption time, 1.5 bar adsorption pressure, 0.1 bar desorption pressure, and a flow rate of 1 SLPM, producing a purity of approximately 94% and a recovery of about 20%. Increasing the adsorption duration negatively affected the oxygen purity but positively influenced process recovery and productivity. Adding an equalization stage improved process recovery by 18.9% for PSA and 14.5% for VSA. Additionally, increased dead volume in the column had adverse effects on purity, productivity, and recovery for both PSA and VSA units.