The pilot tests applied in June 2016 in two contaminated aquifers (Barreiro, Portugal, and La Felguera, Spain) were evaluated in terms of the effect of nanoparticle injection on hydrogeology of aquifers and on removal of heavy metals from groundwater.

In Barreiro it was observed that the nanoparticles showed reduced mobility. Therefore a modification of nanoparticles and injection strategy was proposed. UDE, responsible for nanoparticle production, improved the nanoparticles by reducing the aggregate sizes, and increasing stability time of the colloidal suspensions. The results of slug tests were also used to calibrate the simple 3D numerical model developed by POLITO. This model was then used to predict the nanoparticle transport and their effect on hydrogeology of the aquifer during main injections.

In the case of Spanish site, limited clogging and reduction in hydraulic permeability was observed. The main concern in this case was the heterogeneity of the aquifer. Therefore a more elaborated site characterization was planned. It included, among the other things, permeability measurement at different depths, taking undisturbed cores, etc. These information were carefully analyzed and then implemented into the 3D model developed by POLITO. The calibrated model was then used to predict the main application, providing the consortium with suggestion to improve the efficiency of barrier installment.

Following the remediation plan for test injections, groundwater and soil samples were taken and analyzed before, during, and until four months after each injection. These samples were analyzed by FSU. Analysis of soil samples for Iron provided us with the fate of injection nanoparticles. In Barreiro, most of nanoparticles traveled to shallower portions of aquifer and precipitated near groundwater level and into unsaturated zone. In Nitrastur, the injected nanoparticles followed the most conductive pathways (as groundwater) and resided in such areas. These analyses matched the events and observations during the test injections. Furthermore they suggest that the NPs settled in the close vicinity of the injection points and do not travel with groundwater to unwanted regions.

Post injection groundwater samples were also analyzed and compared to control wells, and pre-injection samples. The analysis of groundwater samples suggested a partial remediation in targeted areas. This partial remediation was found mostly due to heterogeneity of the aquifers. For example, in case of the Spanish site, the injected nanoparticles followed high conductive layers which were placed at the bottom of aquifer, while the source of contamination was within unsaturated zone, where no particles were injected.  Therefore, with each infiltration event, a new plume of heavy metal contaminates may enter the top portion of aquifer. Therefore the location for the barrier for the main application was chosen downstream the pilot area, where the input rate of contaminant was lower. Additional monitoring was also considered to quantify the input of heavy metals from top layers. Despite these heterogeneities, Removal of two main contaminants, As and Zn, was observed in lowest, high conductive portion of the aquifer, through which main part of groundwater flows.

Following a suggestion by Project Advisory Group, a filtration process was added to monitoring and analysis. It was observed that part of heavy metal observed in post-injection groundwater samples (supposedly not removed from groundwater) was actually adsorbed to injected iron oxide nanoparticles that were still mobile in vicinity of sampling points. Therefore a modification in particles and in injection scenarios was applied to ensure the complete precipitation on nanoparticles shortly after injection.

From two pilot injections, we gained more information about the characteristics of each site, as well as the fate of the injected nanoparticles and their efficiency in removal of heavy metals. This information was used to develop the remediation and monitoring plan for the main application.