A reliable methodology for the determination of transport parameters in the design of fixed bed adsorbers

BIDABEHERE, CLAUDIA MARÍA; GARCÍA, JUAN RAFAEL; SEDRAN, ULISES

Buenos Aires

Congreso; 11th World Congress of Chemical Engineering; 2023

AAIQ

Adsorption separation processes are widely used to remove impurities in gaseous or liquid mixtures. The performance of fixed bed adsorbers (breakthrough curves) depends on the mass transfer resistance in the film surrounding the particles, the intraparticle diffusion resistance, the adsorption equilibrium, the surface kinetics, and the bed fluid dynamics. The shape and size of the adsorbent particles affect both the mass transfer resistance in the particles and the flow in the bed. However, in fitting parameters to experimental data, it is usual to assume the flow pattern -frequently unverified- and uniformity in particle shape and size.An appropriate methodology for the experimental determination of system parameters by means of dynamic techniques, which is useful for the design of adsorbers on an industrial scale, is proposed taking parametric sensitivity and correlation into account. The breakthrough curves of a fixed bed adsorber are analyzed by means of a model of n continuous stirred tanks in series, with a linear approximation for the mass balance in the particles, assuming that the characteristic time of the adsorption and diffusion process (Td) involves the resistance to mass transfer (ka)-1 representing the set of particles in the bed, even when they do not have a uniform size [1]. When the relationship (K*) between the concentration of the adsorbed species (Q) in equilibrium with that in the fluid phase (C) is high (K*>>1), which is usual under typical conditions, the flow pattern (represented by n) and the relationship between Td and the mean overall time of the process (Tg), i.e., θd = Td/Tg, determine the efficiency of the bed. If θd < 0.01, the parametric sensitivity is not adequate for the determination of ka. If θd >1 it is possible to determine ka with a low error starting from the initial value of the breakthrough curve, regardless of the degree of axial dispersion in the bed (related to n) [2]. Under this condition (θd >1) it is not necessary to verify the occurrence of plug flow. Systems where data dispersion is high induce large errors in the calculation of Tg and the operation of the adsorber should be modified to provide values of θd ≈ 0.1, in which the parametric sensitivity is adequate. However, in the latter case, the plug flow verification is necessary. An alternative to the typical tests for determining the residence time distributions in fixed bed adsorbers is provided. It is also shown how this method can be applied to systems with non-linear adsorption equilibrium.