AWWA ACE65266

Recent U.S. Environmental Protection Agency (USEPA) regulatory changes have reduced the Maximum ContaminantLevel (MCL) for arsenic from 50 to 10 µg/L. The new arsenic regulations will have the mostdramatic impact on groundwater systems that commonly have minimal wellhead treatmentsystems, and generally have higher arsenic concentrations than surface waters. Arsenic can beremoved by adsorption on metal (hydr)oxides during coagulation or by adsorptive packed-beds,ion exchange resins, and membrane separation systems (McNeill and Edwards 1995; Scott et al.1995; McNeill and Edwards 1997a,b; Brandhuber and Amy 1998, 2001; Chen et al. 2002; Wanget al. 2002; Wingrich and Wolf 2002; DeMarco et al. 2003). Packed-bed adsorption systems areusually cost-effective, have minimal waste streams, and are relatively simple to operate.Although the use of granular oxide media in continuous-flow packed beds is an attractive option forutilities, the actual performance of these technologies can vary by more than a factor of 1000 interms of run length to breakthrough, and currently no basis exists for predicting these variationswithout expensive trial and error testing at pilot scale or at bench scale using rapid small-scalecolumn tests (RSSCTs). Moreover, once the experimental testing phase is completed, the resultsapply only to the exact water tested, and they cannot be used to confidently predict the changesthat will occur in response to even slight changes in water quality (e.g., pH or arsenic speciation).Thus, a robust modeling approach, analogous to what exists for removal of organic chemicals ongranular activated carbon, is needed to complement and extend the available experimentalapproaches, as well as provide a conceptual framework for interpreting performance data. For asuccessful modeling approach, adsorption equilibrium must be described accurately within thematrices of other ions that are encountered in practice. Surface complexation modeling is theappropriate technique for characterizing single and multi-solute adsorption equilibrium insystems of this type. The focus of this research and presentation is on the development ofdiffuse layer model (DLM) and triple layer model (TLM) parameters for arsenate adsorptiononto two commercially available iron based adsorbents (GFH and E33 Bayoxide) in backgroundwaters containing silica and calcium. Macroscopic batch sorption experiments were conducted over a range ofsurface coverage, pH values, and competing ion conditions. DLM and TLM were used tomodel As(V) uptake by the commercial sorbents examined to incorporate the enhancementand/or competition caused by the source water ions. These model parameters will form a selfconsistent database, similar to that developed for ferrihydrite by Dzombak and Morel (1990).The SCMs can then be integrated in a dynamic transport model to predict column life given aparticular source water and to predict the effects of source water changes. The study determinedparameters for the SCMs based on surface titration data, macroscopic batch adsorptionexperiments, and spectroscopic data. The models were calibrated using a limited set of thesingle solute experimental data, and then verified using the remaining data collected over thebroader range of solute and solution conditions. Model calibration was performed usingFITEQL 4.0, a commonly used parameter estimation software package for SCMs. Modelverification was performed using MINEQL+ and/or FITEQL 4.0. Includes 10 references, figures.

Product Details

Edition:

Vol. - No.

Published:

06/01/2007

Number of Pages:

33

File Size:

1 file , 1.9 MB