Date of Award

Summer 7-18-2014

Document Type


Degree Name

Doctor of Philosophy (PhD)


Civil, Environmental & Architectural Engineering

First Advisor

R. Scott Summers

Second Advisor

Karl G. Linden

Third Advisor

Fernando L. Rosario-Ortiz


The removal of organic contaminants in drinking water processes can be accomplished with a variety of technologies. In particular, the use of granular activated carbon (GAC) has been cited as the best available technology for the removal of 51 of 54 regulated organic contaminants (CFR, 1994). The citation of best available technology is due to GAC’s ability to meet today’s regulatory standards. The U.S. Environmental Protection Agency is currently considering reducing the maximum contaminant level for a group of up to 16 carcinogenic volatile organic compounds (cVOC). Additionally, another set of organic compounds, perfluorinated alkyl acids (PFAAs), are attracting regulatory attention due to their ubiquitous presence in the environment and persistent, bioaccumulative and toxic properties. Using a variety of groundwaters, this thesis attempts to address whether GAC is a viable treatment technology to meet lower standards for cVOCs or new standards for PFAAs. Waters containing different background matrices of dissolved organic matter (DOM) spiked with low concentrations of cVOCs (0.1 – 50 μg/L) or already containing trace concentrations of PFAAs (16 - 690 ng/L) were treated with GAC using bench-scale flow-through adsorbers. Scale-up work was accomplished to correlate bench-scale results to full-scale for both groups of compounds.

GAC adsorption capacity for cVOCs was negatively affected by competition in two forms: co-solute competition and DOM competition. Co-solute competition was strongly affected by the similarity in adsorptivites between co-solutes. Grouping co-solutes by their Freundlich adsorption coefficients, co-solutes with similar adsorptivites were found to affect capacity 4-5x more than co-solutes with dissimilar adsorptivites. Output from the Pore and Surface Diffusion Model (PSDM) supported this. DOM negatively affected GAC adsorption capacity for cVOCs to a greater extent. Bed volumes to 10% breakthrough were reduced by an average of 28% when comparing the low-TOC end member water (TOC: 0.3 mg/L) against organic-free water. Larger differences were observed for waters with higher TOC concentrations. Regressions to predict 10% breakthrough were applied to 22 breakthrough curves from four different groundwaters with concentrations of 1,2 dichloroethane (1,2 DCA) spanning 2.5 orders of magnitude. Correlations considered various DOM characteristics as measured by fluorescence, UV spectroscopy, and size exclusion chromatography. The best predictors of bed volumes to 10% breakthrough were Peak C / UV340 * 1,2 DCA concentration (R2 = 0.82, n = 22) and Peak C / UV excitation * 1,2 DCA concentration (R2 = 0.82, n = 22). Avoiding the cost and expertise required for fluorescence analysis, a correlation was developed using only UV254 absorbance and 1,2 DCA concentration (R2 = 0.74, n = 22). The RSSCT over-predicted GAC capacity for cVOCs by a factor of 1.4 – 2.4. The fouling index (FI) proposed by Corwin (Corwin and Summers, 2010) provided mixed results to correct over-capacity predictions of the RSSCT. Two of four FIs obtained using the 95% CI about a regression developed by Kennedy contained the full-scale capacity (Kennedy, 2013). A correlation relating the ratio of the influent target organic concentration to TOC concentration was developed during this effort. Results were satisfying; all four FIs obtained using the 95% CI contained the targeted full-scale capacity breakthrough curve.

The effective use of GAC for the removal of PFAAs from groundwater was demonstrated. GAC efficiency was not affected with increasing EBCT indicative of negligible DOM preloading occurring as a function of adsorber depth. The RSSCT over-predicted full-scale GAC capacity for PFAA by a factor of 1.7. Scale-up efforts used three independent correlations; all regressions overcorrected RSSCT capacities. Variable full-scale influent concentrations, and unequal influent concentrations between the RSSCT and the full-scale adsorber, are believed to be responsible for scale-up difficulty. CURs were calculated for 20 breakthrough curves from both RSSCT and full-scale adsorbers. The 4 carbon-chained PFBA is the only compound where GAC treatment is in the transition range between practical and unfeasible; seven other PFAA compounds met thresholds for feasibility and GAC treatment should be considered practical.