Transcolumn dispersion in a computational mimic of an analytical silica monolith reconstructed from sub-microtomographic scans using computational fluid dynamics

Abstract

Analysis of transcolumn dispersion in an analytical silica monolith is presented via direct numerical simulations in a topological model reconstructed from 3D nanotomographic scans. This was the first instance to incorporate retention in the study of dispersion behavior in reconstructed models for monoliths. The low scanning resolution employed in this work allowed simulating retained behavior with no appreciable loss of prediction accuracy. Heterogeneities of the order of 1-2 domain lengths were observed from peak parking and transient dispersion simulations, indicating a relatively homogeneous core in silica monoliths. Analysis of the short-range interchannel dispersion revealed that equilibration at the domain-level occurred mainly through lateral diffusion. Phenomena not captured by the model, viz. transcolumn eddy dispersion and dispersion due to external film mass transfer resistance, were estimated from deviations of simulated data from experimental values. Three scenarios were explored to model the transcolumn and external film mass transfer resistance contributions to overall dispersion. In the first case, when transcolumn dispersion was controlled by a convection mechanism, the transcolumn velocity biases were within the experimentally observed ranges for silica monoliths. Deviations of estimated film mass transfer resistance from that predicted by penetration theory varied inversely with the zone-retention factor. In the second case, when film penetration theory was assumed to hold good and the transcolumn contribution was modeled to exhibit a coupled diffusive-convective behavior, the corresponding persistence-of-velocity lengths revealed that about 20-30% of the column length was utilized in relaxing the transcolumn concentration gradients. A phenomenological approach was proposed to estimate transcolumn dispersion from the simulated transverse dispersion coefficients and column residence times. The parameters estimated from the proposed approach were compared with the third scenario, wherein transcolumn dispersion was modeled according to Giddings' couple equation, while external film mass transfer resistance was assumed to deviate from film penetration theory. In the absence of any method to predict a priori or directly measure transcolumn dispersion and external film mass transfer resistance in silica monoliths, this approach served as a first approximation to judge the magnitude of each contribution to overall dispersion. © 2013 Elsevier B.V. All rights reserved.