Date of Award

Summer 2023

Document Type

Open Access Thesis


Chemical Engineering

First Advisor

James Ritter


Adsorption-based separations processes, along with the adsorbents that enable them, have benefited from a greater particular focus in recent years, following a desire to improve process energy efficiencies and cost economics. One such adsorbent, carbon molecular sieves (CMS), have likewise been a greater focus. CMS materials offer several key practical uses, such as the separation of nitrogen and oxygen, and the removal of carbon dioxide from methane process streams. In order to effectively design and implement an industrial scale process using a CMS material, the behavior of these gases on the chosen material must be known, including the limiting mass transfer regime. As frequency response methodology has proven useful in the individual determination of limiting mass transfer mechanisms, a volumetric swing frequency response (VSFR) apparatus built in-house at the University of South Carolina was used to collect experimental data, which was then modeled using COMSOL Multiphysics and MATLAB in an effort to identify the characteristic limiting mass transfer mechanisms of nitrogen and oxygen on one such CMS material.

The aforementioned VSFR method is relatively simple in approach: the volume of the system is sinusoidally perturbed, and the differential pressure between this system volume and a reference volume is measured as the response variable to the volume perturbation. The VSFR system constructed at the University of South Carolina is capable of operating at a wide range of frequencies, which allows mass transfer mechanisms to be identified equally as effectively for different gases which may possess a large range of diffusion rates. The frequency response variables, intensity and phase lag, were determined from the data obtained using this VSFR apparatus, and several variations of a micropore model were used to fit this experimental data in an effort to identify the limiting mass transfer characteristics for nitrogen and oxygen on a sample of Shirasagi CMS 3K172. Data was collected at a temperature of 25oC for pressures of 100 and 200 torr, and collected at a pressure of 750 torr at temperatures of 20, 30, 40, and 50oC for both nitrogen and oxygen on this material.

Multiple frequency regions in the experimental data observed for nitrogen and oxygen indicated the presence of several important mechanisms. At low frequencies, large plateau values of the intensity and near-zero values for the phase lag indicated the absence of any limitations to mass transfer, and that the system was mainly governed by local isothermal equilibrium conditions. At intermediate frequencies, a sharp decrease in the experimental intensity curves and the growth of a peak in the phase lag curves signified the emergence of micropore diffusion as the limiting mass transfer mechanism for this frequency range. At high frequencies, the intensity curves leveled off to near-zero values, and the phase lag peaks decreased to low/intermediate value plateaus, which indicated the dominance of resistance at the micropore mouth, as well as some behavior in the macropore that was not approximated by any micropore models.

The best-fit model results for nitrogen and oxygen on Shirasagi CMS 3K172 indicated that the most likely mass transfer model for nitrogen on this material was a single site micropore diffusion limitation paired with a limitation at the micropore mouth, and that both of these mass transfer coefficients followed a more complex loading dependence than Darken’s original model. This behavior was described by an empirical Qinglin’s loading dependence approach. For oxygen, the most likely mass transfer model was observed to be a dual site micropore diffusion limitation paired with a single micropore mouth limitation. Likewise, each of these three mass transfer coefficients were observed to follow a loading dependence more complex than that which is predicted by Darken’s model.