Dissolved organic matter adsorption from surface waters by granular composites versus granular activated carbon...

Many new sorbents have been introduced as an alternative for granular activated carbon (GAC), the most common sorbent for dissolved organic matter (DOM) removal. In the current study, we developed an applicable granular composite based on a flocculant commonly employed for drinking water treatment adsorbed to montmorillonite. DOM adsorption from surface waters, Lake Kinneret and Suwannee River, with low and high specific ultraviolet absorption (SUVA), respectively, by composite and GAC columns, was studied. Adsorption of DOM from Suwannee River was significantly higher by the composite column, in comparison to the GAC column, while an opposite trend was obtained for the adsorption of DOM from Lake Kinneret. In-situ regeneration of the columns with a brine solution was extremely efficient and inefficient for the composite and GAC columns, respectively. Adsorption, of both waters, post-regeneration by the composite column was not compromised, while GAC effectiveness decreased. The opposite trend in DOM adsorption from Suwannee River and Lake Kinneret was explained by the different affinities of the sorbents towards various DOM molecules. Distinguishing between different DOM components adsorbed by GAC and the composite was supported by ¹³C NMR and direct pyrolysis-GC-MS measurements. Furthermore, we demonstrated that the kinetics and adsorption at the equilibrium of five organic molecules to the composite and GAC can be correlated to their chemical-physical properties. Indeed, combining the properties of both sorbents, by integrating them into a single column, yielded higher DOM removal than by the individual columns. Furthermore, since DOM removal by GAC and by the composite, increases, and decreases with temperature, respectively, the integrated column, mitigates the changes in removal, stabilizing the adsorption performance. Such an integrated filter may minimize additional seasonal and water quality fluctuations.