Competitive adsorption models for the dimensioning of activated carbon filters

In groundwater treatment plant (GWTP) adsorption is often the last treatment step to remove recalcitrant compounds from contaminated water in order to reach the groundwater threshold values. Granular Activated Carbon (GAC) is the most common adsorbent material since it combines high efficiency with low cost. Generally, the design of adsorption stage is based on the isotherm of each compound. However, sometimes, a complex mix of contaminants is present in the water to treat. In these cases, a competitive effect can arise from the multicompound interaction with the GAC: they compete with each other to occupy the GAC available sites. As results of this phenomenon, the operating life of the filter can be reduced compare to the designed one. The aim of the thesis is to study competitive adsorption and to simulate the contaminant breakthrough from an adsorption filter. This work is focused on competitive interactions of three chlorinated solvents, often presented in contaminated aquifers: Trichloroethylene (TCE), 1,2-Dichloroethane (1,2-DCA) and cis-1,2-Dichloroethylene (c-1,2-DCE). Firstly, the affinity between these compounds and GAC was studied in non-competitive conditions, the single adsorption isotherms were estimated from experimental batch data. Freundlich model (E.Worch,2012) was used for all the three tested substances. It emerged that TCE is the compound which has higher affinity, followed by c-1,2-DCE and 1,2-DCA.Non-competitive Freundlich model, used for all the investigated compounds, overestimates the experimental data obtained from competitive batch tests. Therefore, competitive adsorption models were studied to fill this gap. In particular, in this work, Ideal Adsorbed Solution theory (IAST) model and two simplified ideal adsorbed solution (SIAS) models were used(E.U. Schlonder, W. Fritz,1980). The aim of this thesis is to understand which competitive model is better to fit competitive batch test and easier to use in GAC filter design. The main parameter to design a filter is the operating life. Indeed, once the filter reaches contamination threshold value it becomes exhausted and need to be regenerated or replaced. The model simulations, performed in the last part of the work, allow the estimation of the operating life of the filter, emphasising the differences competitive and non-competitive conditions. As results from not competitive and competitive batch test and investigated models, it is noted that the three chlorinated solvents TCE, 1,2-DCA and c-1,2-DCE in the same mix show a competitive behaviour, and in particular TCE has higher affinity that causes a reduction of 1,2-DCA and c-1,2-DCE adsorption. The implementation of different models, and the study of error with respect to the experimental data, suggest that the better competitive model, in terms of accuracy and simplicity, is the SIAS model which uses the same n Freundlich parameter. So that, this model is implemented to perform the simulations. The simulation predicts the behaviour of a standard-sized filter. During competition, contaminants leave the filter in reverse affinity order: 1,2-DCA is always the first contaminant to leave the filter, displaced by contaminants more akin to activated carbon. This results in a reduction of the operating life of the filter in competitive conditions. It is estimated that the reduction of the filter in competitive conditions ranges between 5 to 60%.