Colloid transport processes: Experimental evidence from the pore scale to the field scale

Abstract

Aquatic colloids play a central role in mediating the transport of contaminants in subsurface media. Depending on the environmental conditions, the transport of contaminants can be enhanced or retarded in the presence of colloids. The changes of the transport velocities and mass transfer rates of colloid-associated contaminants are of high environmental relevance. The keys to the assessment of the effects of colloids on subsurface transport is the colloid-contaminant interaction and the mobility of the colloid itself. While the theory of colloid transport was laid out in the late 1950's, the actual parametrization of colloid transport has long been concealed by the used experimental conditions. Classical column tests obscure the pore scale processes by limiting the access to pore scale heterogeneities. Recently the transport of colloids and contaminants was unveiled using magnetic resonance imaging (MRI) and micromodel experiments. MRI provides non-invasive access to dynamic processes in porous and fractured media with a spatial resolution down to 200×200×200 μm3 and a temporal resolution in the seconds range. MRI data were used to characterize the pore space, to measure the flow and diffusion of water, to quantify the attachment of colloids and the transport of contaminants in porous media. The transport of water, certain metal ions and several colloids can be measured without additional tracers. Other potential contaminants are accessible using magnetic tagging, or indirectly by their effect on the relaxation times of water. The main limitations of MRI include susceptibility artefacts and the dependence of the MRI signal on a variety of boundary conditions, sometimes leading to ambiguous calibration curves. Micromodels, i.e. porous structures etched in silicon wafers, glass, or other materials, provide single-particle-single-pore access. With a spatial resolution below 1 μm and a time resolution in the millisecond range, the processes at the pore scale are accessible. The results demonstrate that the theoretical framework of colloid filtration can be applied to the pore scale transport of colloids. It was shown, that the main causes for deviations of experimental results from theoretical predictions are particle-particle interactions, inaccurate descriptions of the flow field, chemically heterogeneous surfaces and additional collector surfaces. Air bubbles and other nonmiscible liquids can be considered as temporary collectors, since the colloids attached to the air-waterinterface are released when the air bubble dissolves. Non-polar interfaces cause a repartitioning of non-polar contaminants, thus limiting the effects of colloid transport. While MRI and micromodels can significantly enhance our understanding of colloid transport phenomena, upscaling to field conditions is still difficult and requires an extensive and detailed data set. Laboratory, pilot scale and field investigations provide a quick way to obtain site and situation specific data on colloid transport processes. In column tests, the attachment probability, summarizing particle-collector and to some extent particle-particle interactions, can be parametrized using the ionic strength and the dominating cation of the solution. Column tests illustrated that the transport of colloids is scale dependent. Since the typical length scale of column tests is well below the theoretical transport distance, column tests tend to overpredict colloid transport. Changes of the hydrodynamic and hydrochemical conditions can cause significant colloid mobilization. This effect was also observed at the pilot scale and at the field scale. Pilot and field scale experiments showed that the transport of metal ions is only very little affected in the presence of colloids. At transport distances of 200 m in a calcareous gravel aquifer, about 10% of the metal ions were found associated to colloids. Since the transport distance of the colloids themselves is far less, a dynamic equilibrium between colloids and dissolved metal ions was proposed. With increasing transport distance, a slight increase of colloid-associated metal ions has been found. The results limit the importance of colloid transport in the investigated shallow aquifers to otherwise immobile contaminants like PAHs or radionuclides. One scenario, where several of the preconditions for colloid transport seemed to be fulfilled, is the emission of contaminants from landfills with direct contact to groundwater. Here, a high concentration of colloids and high concentrations of contaminants come together. Our research shows, that the interface between the landfill body and the groundwater is an effective hydrochemical barrier for colloids. The colloids downgradient of the disposal sites and the colloids inside the disposal differ significantly. We propose a self-sealing effect of the landfill, caused by landfill colloids being filtered at the interface, and by this reducing the hydraulic conductivity of the interface. Field investigations verified this hypothesis. More over, detailed investigations of the colloid-metal associations revealed that the highest concentrations of metal ions were found with particles greater than 10 μm, i.e. the least mobile particles.