nZVI mobility and transport: laboratory test and numerical model

Zerovalent iron nanoparticles (nZVI) are becoming one of the most widely recommended nanomaterials for soil and groundwater remediation. However, when nZVI are injected in the groundwater flow, the behavior (mobility, dispersion, distribution) is practically unknown. This fact generally results in the use of enormous quantities of them at the field scale. The uncertainties are on the effective volumes reached from the plume of nZVI because their tendency to aggregate and their weight can cause their settling and deposition. So, the mobility of nanoparticles is a real issue, which can often lead to inefficient or expensive soil remediation. Furthermore, there is another aspect that must be considered: the fate of these nZVI in the groundwater and their possible impact on the subsoil environment. All these considerations have led us to propose an application of nZVI simulating the permeation technique through a laboratory experience, finalized to have a better, or even simpler description of their real behavior when injected in a flow in the subsoil. A two-dimensional laboratory-scale tank was used to study the dispersion and transport of nZVI. A nZVI solution, with a concentration equal to 4.54 g/L, was injected into glass beads, utilized as porous medium. The laboratory experiment included a digital camera to acquire the images. The images were then used for calibrating a numerical model. The results of the mass balance confirm the validity of the proposed numerical model, obtaining values of velocity (5.41 x 10(-3) m/s) and mass (1.9 g) of the nZVI of the same order of those from the experimental tests. Several information were inferred from both experimental and numerical tests. Both demonstrate that nZVI plume does not behave as a solute dissolved in water, but as a mass showing its own mobility ruled mainly from the buoyancy force. A simple simulation of a tracer input and a nZVI plume are compared to evidence the large differences between their evolution in time and space. This means that commercial numerical models, if not corrected, cannot furnish a real forecast of the volume of influence of the injected nZVI. Further deductions can be found from the images and confirmed by means the numerical model where the detachment effect is much smaller than the attachment one (ratio k(d)/k(a) = 0.001). From what is reported, it is worthwhile to pay attention on the localization of the contaminants source/plume to reach an effective treatment and it is important to go further in the improvement of solution for the limiting the nanoparticles aggregation phenomenon.