Competitive Effects of Natural Organic Matter: Parametrization and Verification of the Three-Component Adsorption Model COMPSORB LI DING, BENITO J. MARIN ˜ AS,* LANCE C. SCHIDEMAN, AND VERNON L. SNOEYINK Department of Civil & Environmental Engineering and Center of Advanced Materials for the Purification of Water with Systems, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801 QILIN LI Department of Civil, Construction & Environmental Engineering, Oregon State University, Corvallis, Oregon 97331 Natural organic matter (NOM) hinders adsorption of trace organic compounds on powdered activated carbon (PAC) via two dominant mechanisms: direct site competition and pore blockage. COMPSORB, a three-component model that incorporates these two competitive mechanisms, was developed in a previous study to describe the removal of trace contaminants in continuous-flow hybrid PAC adsorption/membrane filtration systems. Synthetic solutions containing two model compounds as surrogates for NOM were used in the original study to elucidate competitive effects and to verify the model. In the present study, a quantitative method to characterize the components of NOM that are responsible for competitive adsorption effects in natural water was developed to extend the application of COMPSORB to natural water systems. Using batch adsorption data, NOM was differentiated into two fictive fractions, representing the strongly competing and pore blocking components, and each was treated as a single compound. The equilibrium and kinetic parameters for these fictive compounds were calculated using simplified adsorption models. This parametrization procedure was carried out on two different natural waters, and the model was verified with experimental data obtained for atrazine removal from natural water in a PAC/membrane system. The model predicted the system performance reasonably well and highlighted the importance of considering both direct site competition and pore blockage effects of NOM in modeling these systems. Introduction Competition between target trace compounds and natural organic matter (NOM) is common in activated carbon processes used for drinking water treatment. NOM is a complex, site-specific mixture of numerous organic com- pounds present in both ground and surface waters at concentrations of 0.5-12 mg/L as carbon (1), having mo- lecular weights (MW) of several hundred to over 10 000. NOM has been shown to reduce the adsorption capacity of activated carbon for trace compounds, typically present at concentra- tions in the microgram- or nanogram-per-liter level, when both substances are exposed to the adsorbent at the same time, and when carbon is equilibrated with NOM prior to being exposed to the trace compound (2, 3). Carter et al. identified direct site competition and pore blockage as the two primary mechanisms of competitive adsorption between trace compounds and NOM (4). Small, strongly adsorbing NOM molecules with a size comparable to that of the target compound adsorb in the same size pores as the target compound, thus exerting direct site competition and causing a reduction in adsorption capacity for the target compound. In contrast, large NOM molecules that may not adsorb on the same sites as the target compound are capable of constricting or blocking pores, through which the trace compound travels to access final adsorption sites. This phenomenon reduces the rate of adsorption kinetics for the trace compound (5), and the extent of this effect depends on molecular weight distribution (MWD) of the NOM as well as the pore size distribution and configuration of activated carbon pores. Li et al. found evidence that the pore-blocking fraction of NOM is mostly in the MW range of 200-700 dalton (6). Powdered activated carbon (PAC) can be effectively applied to water, together with the membrane filtration process, in reactor configurations such as hybrid PAC adsorption/membrane filtration systems (7). These hybrid systems make it possible to simultaneously remove dissolved and particulate contaminants by adsorption and filtration mechanisms, and they allow for carbon contact times much longer than the reactor hydraulic retention time, which enables more efficient use of PAC adsorption capacity. However, the typical carbon retention times used in these hybrid systems (less than 2 h) are usually still inadequate to achieve adsorption equilibrium. Li et al. (8, 9) made significant progress in the study of competitive adsorption in flow-through systems by develop- ing the three-component model COMPSORB and applying it to quantify the effects of competition between a trace target contaminant and two model compounds representing the strongly competing (SC) and pore blocking (PB) fractions of NOM. This model was able to accurately predict the rate of adsorption for all three compounds. The primary objective of this research was to develop a procedure to extend the application of the COMPSORB model to natural water systems where a variety of unknown compounds in the background NOM contribute to competi- tive adsorption effects by direct site competition and pore blockage. The model was then verified by comparing model predictions with experimental data obtained with a continu- ous-flow PAC/membrane system. Materials and Methods Water. Distilled deionized (DDI) water as organic free water (OFW) and two natural waters were used in this study. Groundwater from the Clinton Water Works (CWW), Clinton, IL, was collected before treatment and stored in a stainless steel barrel at 4 °C in a cold room. Prior to an experiment, the water was warmed to ambient temperature and passed through a nylon membrane filter with a nominal pore size of 0.45 µm (OSMONICS, Minnetonka, MN) to remove suspended solids. The dissolved organic carbon (DOC) concentration of CWW water, measured using a Phoenix 8000 * Corresponding author phone: +1 217 333 6961; fax: 1 217 333 6968; e-mail: marinas@uiuc.edu. Environ. Sci. Technol. 2006, 40, 350-356 350 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 1, 2006 10.1021/es050409u CCC: $33.50  2006 American Chemical Society Published on Web 12/02/2005