No abstract

The role of granular activated carbon (GAC) surface chemistry on the adsorption of four model dissolved organic material (DOM) isolates and four surface water natural organic material (NOM) samples was investigated by using () ten carbons prepared by modifying the surfaces of a coal-based and a wood-based carbon and () seven different as-received GACs. Because changes in the pore structure resulting from surface treatment were small, while changes in the surface chemistry were significant, the impact on the DOM and NOM uptake by surface-treated carbons was systematically linked to the changes in the carbon surface chemistry alone. For the surface-treated coal-based carbons, there was () no significant capacity difference between acid-washed and heat-treated carbon samples, () oxidation of the carbon surface significantly decreased the uptake, and () the capacity was partially restored by subsequent heat treatment of the oxidized surfaces. A decreasing uptake with increasing surface acidity was evident, and the effects of surface acidity on uptake were qualitatively similar to the two SOCs studied in Part 1 of this series. The experiments with as-received coal-based carbons exhibited the same behavior; however, the reactivity of modified and as-received carbons for DOM and NOM uptake was significantly different. For the wood-based carbon, the impact of surface treatment on adsorption of DOMs and NOM was surprisingly minimal or absent. This finding was in contrast to the effects of surface acidity on uptake of the two SOCs studied in Part 1 in this series. Overall, the reactivity of carbon surfaces to DOM and NOM uptake depended on the raw material type, activation conditions and surface treatment.

Naturally occurring, macromolecular dissolved organic matter (NOM) is known to foul activated carbon adsorbents, reducing the ability of fixed-bed adsorbers to efficiently remove targeted synthetic organic contaminants (SOCs). An accurate description of the effects of NOM competition on SOC adsorption equilibria is required to develop dynamic models, which have application to process design and analysis. A model was developed, using an approach based on the Ideal Adsorbed Solution Theory (IAST), to predict trichloroethylene (TCE) adsorption by activated carbon preloaded with humic acid. The IAST model was formulated for a bisolute system in which TCE and humic acid single-solute uptakes were described by the Langmuir−Freundlich and Freundlich isotherms, respectively. The humic mixture was modeled as a single component based on previous studies that identified the low-molecular-weight hydrophobic fraction as the most reactive with regard to preloading effects. Isotherms for this fraction, isolated from whole humic acid using ultrafiltration, were measured, and molar concentrations were computed based on an average molecular weight determined using size-exclusion chromatography. The IAST model was modified to reflect the hypothesis that TCE molecules can access adsorption sites which humic molecules cannot and that no competition can occur on these sites. The model was calibrated with data for TCE uptake by carbon preloaded with the low-molecular-weight humic acid fraction and was verified by predicting TCE uptake by carbon preloaded with whole humic acid. Further improvement to the model was possible by accounting for pore blockage as a mechanism which can reduce the effective surface area available to TCE.

Five natural waters with a broad range of DOC concentrations were fractionated using various coal- and wood-based granular activated carbons (GAC) and alum coagulation. Adsorption and alum coagulation fractionated NOM solutions by preferentially removing components having high specific ultraviolet absorbance (SUVA). UV absorbing fractions of NOM were found to be the major contributors to DBP formation. SUVA appears to be an accurate predictor of reactivity with chlorine in terms of DBP yield; however, it was also found that low-SUVA components of NOM have higher bromine incorporation. SUVA has promise as a parameter for on-line monitoring and control of DBP formation in practical applications; however, the effects of bromide concentration may also need to be considered. Understanding how reactivity is correlated to SUVA may allow utilities to optimize the degree of treatment required to comply with DBP regulations. The reactive components that require removal, and the degree of treatment necessary to accomplish this removal, may be directly obtained from the relationship between SUVA removal and the degree of treatment (e.g., alum dose).

No abstract