Formatted Title
Expanding the Activated Carbon Story with New Data: A Column Study to Empirically Validate Sediment Cap Modeling Predictions
Background/Objectives
Contaminated sediment remedy designs utilize the adsorption benefit of activated carbon (AC) for hydrophobic organic contaminants (HOCs), such as polychlorinated biphenyls (PCBs) and polycyclic aromatic hydrocarbons (PAHs), to increase the lifetime of the remedy for various performance benchmarks, most notably, the porewater concentrations of the HOC at the top of the cap. The use of numerical simulation models, such as CapSim, have been widely adopted for predicting the performance of these forms of AC based sediment remedies. For this reason, it is important to understand how AC data are being incorporated into this modelling approach and what AC characteristics may impact the output from the model. While several lab and field studies have demonstrated the performance benefits of smaller particle sizes, these benefits are not always apparent in the results of modeling and design approaches using AC in sediment remedies. This tends to support a widely held notion that although granular forms of AC (GAC) experiences lower kinetics, the long-term adsorption behavior will equal powdered AC (PAC). This has also resulted in an expected performance equivalence between PAC and GAC for sediment remediation, even in areas that experience high groundwater upwelling.
The authors have recently presented data from a unique 50-wk study evaluating PCB adsorption by PAC and GAC. The study showed (1) significant order-of-magnitude improvement in adsorption speed and capacity at the smaller particle size, and (2) a large potential for further performance degradation due to nonuniform mixing and separation during subaqueous placement of mixed sand/GAC material. However, this study evaluated these effects in a batch/jar environment and therefore the effect of the slower kinetics of GAC have not been demonstrated in a controlled dynamic environment (lab or field). While the batch study provided strong evidence against the notion of long-term equivalence, there is still a need to demonstrate the effect of the slower GAC kinetics in a flowthrough environment. Results of the flowthrough test can then be compared to a representative CapSim model to identify model conditions and assumptions that create a divergence between empirical and predicted results.
The goal of this ongoing work is to provide critical knowledge on existing issues and answer the question, “Are the current modeling approaches providing accurate/representative outcomes for GAC-Sand mixed caps in dynamic sediment environments with high levels of groundwater upwelling?”
Approach/Activities
The kinetics effect of GAC in a flowthrough environment was studied by setting up three columns that were run side by side for 6 months at an accelerated contaminant transport rate in order to accelerate breakthrough. Column 1 consisted of ‘Sand – No AC’ as a negative control. Column 2 consisted of ‘Sand uniformly mixed with GAC’ to match the CapSim assumption of ideal mixing. Column 3 consisted of ‘Sand nonuniformly mixed with GAC’ to match the field observed particle separation during subaqueous placement. A CapSim model representing the column scenarios evaluated whether the model could reliably predict the outcomes for each scenario. Modifications to the CapSim model were conducted to identify what model conditions and assumptions are necessary to incorporate observed GAC kinetic effects.
Results/Lessons Learned
AC is an important part of the remedy design for contaminated sediment sites, however, the treatability testing and modelling methodology can compromise the expected performance of the remedy, especially when implications of the form of AC (PAC or GAC) are not considered. The particle size of the AC results in very significant differences in the effective equilibrium, which will be reflected in the ability of the remedy to prevent contaminant breakthrough for the given design consideration.
Based on the data from the previous 50-wk adsorption study, it is expected that the GAC/Sand column will experience rapid breakthrough, likely before the unmodified CapSim model would predict. It is anticipated that modifications to the CapSim model will provide results that will provide improved prediction of effluent concentrations. These modifications are based on data from the previous 50-wk study and will thereby provide a more accurate representation of the realistic conditions of (1) nonuniform mixing of GAC and sand within CapSim and (2) incorporating reduced adsorption due to slower GAC kinetics.