The effective radial thermal conductivity (k eff ) of a 2-D analog of a 3-D, parallel channel, corrugated metal, structured adsorbent bed was studied using COMSOL Multiphysics. This 2-D structure consisted of alternating sections of corrugated and flat metal foil sheets, with k eff predicted in 1-D perpendicular to the flat metal foil sheet, i.e., the radial direction in a 3-D cylindrical bed. The effect of the thickness of zeolite coating, thickness of metal, type of metal, type of contact between the metal foil sheets (i.e., metal-to-metal, coating-to-coating and metal-to-coating point contacts, and metal-to-metal imbedded contacts), air gap size between the corrugated and flat metal foil sheets, coating on just one or both sides of the metal foil sheets, alignment of the corrugation between sections, and type of stagnant gas medium on k eff was studied. In all cases, temperature contour plots revealed the minute region around the point contacts, being mostly stagnant gas medium, manifested a significant resistance to thermal conductivity, with the imbedded contacts minimizing the effect. The parametric study revealed direct metal-to-metal contact had the most positive effect on k eff , whether being a point or imbedded contact: k eff respectively varied between 0.561 and 0.726 W m −1 K −1 for SS and 6.66 W m −1 K −1 for Al, showing strong dependence on the metal conductivity and weak dependence on the gas medium and all other parameters. When the corrugated and flat metal foil sheets were either coated with zeolite or separated by an air gap, k eff was significantly reduced, varying between 0.090 and 0.125 W m −1 K −1 in air for SS or Al; k eff also depended strongly on the gas medium but only weakly on the metal conductivity and all other parameters.

The new kinetically limited linear driving force (KLLDF) model was assessed against the traditional LDF model in the prediction of twelve different ternary and quaternary experimental breakthrough curves. These breakthrough curves comprised mixtures of CO2, N2 and CH4 in He adsorbed on carbon molecular sieve MSC 3K 172 and were conducted at various pressures (30, 50 and 100 psia) and at ambient temperature. The LDF and KLLDF models were implemented in the dynamic adsorption process simulator (DAPS) with the loading dependent LDF mass transfer coefficients and the single gas equilibrium adsorption isotherms measured independently with gravimetric uptake experiments. To make the comparison between the LDF and the KLLDF models as fair as possible, they utilized the same set of thermodynamic and kinetic parameters in DAPS, with no adjustments to any of them. Both the LDF and KLLDF models provided reasonable predictions of the experimental breakthrough curves and in-bed temperature histories, with general trends of no CH4 uptake, gradual N2 uptake and fast CO2 uptake. However, the KLLDF model always provided better predictions, especially when CO2 was present. The results revealed that the traditional LDF model led to depressed adsorbed phase loadings of CO2, thereby underpredicting its breakthrough time in all cases. This depression stemmed from the equilibrium loading in the LDF driving force of the LDF model depending on the gas phase partial pressure of each component outside the pore structure. In contrast, the KLLDF model corrects this issue by making the equilibrium loading in its LDF driving force dependent on the actual loading of each component inside the pore structure. In conjunction with the extended mixed gas Langmuir model, the KLLDF model is perhaps the more appropriate model to use instead of the LDF model for any multicomponent adsorbate-adsorbent systems, whether diffusion limited or not, since it reduces to the LDF model for systems that do not exhibit significant diffusional limitations.