BENCH-SCALE TESTS AND MODELING OF ADSORPTION OF NATURAL ORGANIC-MATTER BY ACTIVATED CARBON

Adsorption equilibria for activated carbon removal of natural organic matter (NOM) were obtained for a lake water, a leaf extract humic material and a sample from a monitoring well serving a municipal solid waste landfill. A multicomponent model incorporating ideal adsorbed solution theory (IAST) and a non-adsorbable fraction more accurately described equilibrium data for the overall isotherm than a Freundlich single component isotherm model, and adequately simulated isotherms for variable initial concentration of the water as represented by a lumped parameter. A methodology is described for estimating equilibrium parameters for multiple hypothetic components that requires fewer parameter searches and is thus less computationally demanding than some other modeling schemes using IAST. The acquired equilibrium information was subsequently used to calibrate a dynamic, fixed-bed model in accordance with bench scale column data for the lake water. Rate parameters evaluated therein were used to predict breakthrough profiles in bench-scale columns of greater bed depth. Generally speaking, dynamic modeling which incorporated the multicomponent equilibrium approach provided more accurate description and prediction of NOM breakthrough profiles than a single component modeling scheme. The multicomponent approach was better able to simulate both the slow approach to exhaustion in the bed and the effects of changes in influent concentration. Although not fully representative of a full-scale system operating over a lengthy time period, the exercise gives insight into dynamic and modeling phenomena which may accompany seasonal variations in natural water quality.