An experimental pressure swing adsorption unit has been constructed and used to investigate the separation of carbon dioxide from carbon dioxide enriched air using both an activated carbon and a type 5A molecular sieve adsorbent. Continuous, cyclic operation was achievedusing a pair of fixed bed adsorbers. At any one time the feed gas entered one bed at a high pressure and part of the purified gas was returned to the other bed at a reduced pressure to provide countercurrent regeneration of the adsorbent. The beds of adsorbent used were each nominally 0.165m diameter and Im. deep. Separations were carried out at approximately ambient temperature using air flow rates in the range 0.15 to 0.95 kg/m2s and inlet carbon dioxide concentrations'in the range 0.1 to 1.5% v/v. Adsorption pressures of 2 to 6.4 bar were examined, the desorption pressure being maintained throughout at essentially 1.0 bar. The period time was varied from 30 to 900 seconds and the revert ratio (i. e. the ratio of the product gas returned for desorption to the total feed rate to the unit) was varied from 0 to 1.0. The carbon dioxide separation efficiency was found to increase markedly as the adsorption pressure and the revert ratio were increased whereas it was relatively insensitive to variations in feed rate and, more particularly, feed concentration. The performance of the molecular sieve adsorbent was found to be very sensitive to the presence of moisture in the feed gas. In contrast the carbon dioxide efficiencies observed with Lhe activated carbon were unaffected by the presence of small amounts (circa 100 ppm) of moisture in the feed. A theoretical model has been proposed for predicting the performance of pressure swing adsorption systems of the type investigated and approximate analytical equations and more precise numerical techniques have been established to represent its solution. The approximate analytical solutions were found to give close agreement with the more precise methods examined under conditions corresponding to low values of a dimensionless period time parameter. The proposed theoretical model incorporates an effective irean mass transfer coefficient to represent the diffusion process within the adsorbent particles. Methods for estimation of the value of this coefficient based on the limiting conditions of a periodic constant surface flux or a periodic constant surface concentration are presented. The experimental performance data were analysed in terms of the proposed analytical solution to give values of the apparent solid phase mass transfer coefficient for comparison with those predicted theoretically. In general the apparent experimental values were consistently less than the predicted values. In addition the relationship between the experimental and predicted coefficients was found to be dependent on both the nature of the adsorbent and a parameter formed by the product of the revert ratio and the adsorption to desorption pressure ratio. Empirical correlating equations which incorporate this dependence are presented.

An experimental pressure swing adsorption unit has been constructed and used to investigate the separation of carbon dioxide from carbon dioxide enriched air using both an activated carbon and a type 5A molecular sieve adsorbent. Continuous, cyclic operation was achievedusing a pair of fixed bed adsorbers. At any one time the feed gas entered one bed at a high pressure and part of the purified gas was returned to the other bed at a reduced pressure to provide countercurrent regeneration of the adsorbent. The beds of adsorbent used were each nominally 0.165m diameter and Im. deep. Separations were carried out at approximately ambient temperature using air flow rates in the range 0.15 to 0.95 kg/m2s and inlet carbon dioxide concentrations'in the range 0.1 to 1.5% v/v. Adsorption pressures of 2 to 6.4 bar were examined, the desorption pressure being maintained throughout at essentially 1.0 bar. The period time was varied from 30 to 900 seconds and the revert ratio (i. e. the ratio of the product gas returned for desorption to the total feed rate to the unit) was varied from 0 to 1.0. The carbon dioxide separation efficiency was found to increase markedly as the adsorption pressure and the revert ratio were increased whereas it was relatively insensitive to variations in feed rate and, more particularly, feed concentration. The performance of the molecular sieve adsorbent was found to be very sensitive to the presence of moisture in the feed gas. In contrast the carbon dioxide efficiencies observed with Lhe activated carbon were unaffected by the presence of small amounts (circa 100 ppm) of moisture in the feed. A theoretical model has been proposed for predicting the performance of pressure swing adsorption systems of the type investigated and approximate analytical equations and more precise numerical techniques have been established to represent its solution. The approximate analytical solutions were found to give close agreement with the more precise methods examined under conditions corresponding to low values of a dimensionless period time parameter. The proposed theoretical model incorporates an effective irean mass transfer coefficient to represent the diffusion process within the adsorbent particles. Methods for estimation of the value of this coefficient based on the limiting conditions of a periodic constant surface flux or a periodic constant surface concentration are presented. The experimental performance data were analysed in terms of the proposed analytical solution to give values of the apparent solid phase mass transfer coefficient for comparison with those predicted theoretically. In general the apparent experimental values were consistently less than the predicted values. In addition the relationship between the experimental and predicted coefficients was found to be dependent on both the nature of the adsorbent and a parameter formed by the product of the revert ratio and the adsorption to desorption pressure ratio. Empirical correlating equations which incorporate this dependence are presented.