Date of Award


Document Type

Open Access Dissertation


Chemical Engineering


College of Engineering and Computing

First Advisor

James A. Ritter


First adsorption equilibrium isotherms for N2, O2, Ar and CO2 on zeolite 13X were measured at three temperatures (25, 50, 75 °C) and pressures up to 110 kPa with a volumetric method and validated by also measuring the same isotherms with a gravimetric system.

Then a comprehensive 2-D mathematical model was developed to study the fundamentals of adsorption processes in parallel channel structured adsorbents. The results clearly indicated that under studied circumstances the plug flow condition can be assumed and therefore the system is not subject to premature breakthrough.

The pressure drop of corrugated structured adsorbents with narrow triangular channels was then investigated both experimentally and numerically. A 1-D model in the form of Darcy-Weisbach equation was developed and successfully tested for different components under different density and viscosity conditions.

Next DAPS equipped with 1-D pressure drop expression was used in preliminary studies proving the capability of structured adsorbent in achieving the desired performance with bulk densities as low as 240 kg/m3. Then a complex single bed setup was equipped with the structured adsorbents and was analyzed with breakthrough runs revealing the average adsorbent layer thickness and the significant dispersion in the system. A set of experiments tested the designed PSA cycle with the newly developed adsorbent. Promising results for recovery (> 93.0 vol% in all cases except one) was obtained. Experimental results were validated with accuracy using DAPS concluding that the adsorbent bulk density was too low to produce high purity CO2 and that by increasing the adsorbent layer thickness this issue can be resolved. Then a step by step scale up procedure was followed using validated DAPS. Results indicated that although structured adsorbents are an excellent choice for high pressure applications, they encounter inherent limitations under vacuum conditions leading to insufficient regeneration during CnD and LR steps and a lowered performance. It was shown that reducing the bed size by increasing the number of the units along with increasing the regeneration time by increasing the total cycle time resolved the problem and the desired performance was achieved.