(A) study on hydrogen purification by pressure swing adsorption process using double layered adsorption column

Abstract

Hydrogen purification from the reformer off gas by the PSA process using the double layered beds of activated carbon and zeolite 5A was studied. The adsorbents used in this study were selected from the two kinds of activated carbon and three kinds of zeolite 5A. The pure, binary and ternary adsorption equilibria of the main impurities and hydrogen were measured on those adsorbents and the isotherm model accurately fitting the experimental data was determined among the known adsorption equilibrium models. The adsorber dynamics in the double layered beds was studied through the breakthriugh experiments and the effects of the activated carbon layer height on the bed utilization were investigated with the mathematical model. The effects of the two important operating variables, the amount of purge and the cocurrent depressurization pressure, on the process performances of the 5-step PSA cycles which include all the characteristic steps of the $H_2$ PSA process were analyzed with the idealized model-equilibrium model. The optimum activated carbon layer heights in the 4-bed $H_2$ PSA process using the double layered beds were determined experimentally and theoretically. The pure component adsorption equilibrium data were fitted with the known adsorption isotherm models - Langmuir, Langmuir-Freundlich, Toth and O``Brien-Myers isotherm models. As expected, the isotherm models that have three adjustable parameters like L-F and Toth isotherm models more accurately predicted the adsorption equilibrium. Binary and ternary adsorption equilibrium data on the activated carbon and zeolite were predicted by the EL, LRC and IAS theory since the models allow fast equilibrium calculation, thus, can be easily incorporated to the PSA process simulation. The IAS theory requires the pure component isotherm to predict the multicomponent equilibrium. In this study, the Langmuir, O``Brien-Myers, L-F and Toth isotherms were used. The IAS theory successfully predicted the experimental data...