The way in which colloidal particles are transported and eventually retained throughout the subsurface porous media is a multi-scale problem that has direct implications for environmental, agricultural and public health sectors. Classic and current filtration models fail to accurately predict the fate of colloids, because they are largely based on mean-field values of surface and flow properties or because they do not account for the complexity of real systems.
The goal of this research is to improve predictive models and describe important mechanisms governing the fate and transport of colloids in the subsoil across different spatial scales--interface scale, pore/collector scale, Darcy scale--with a focus on silver engineered colloids and conditions relevant to agricultural soils. The work comprises a combination of novel experimental techniques and modeling approaches to improve our understanding of: i) short range forces affecting the attachment of colloids to surfaces in complex solution chemistries, ii) pore-scale locations where colloids are statistically most likely to be retained, iii) the influence of shear-induced aggregation on the overall particle distribution and re-entrainment into the bulk liquid phase, and iv) the link between the mechanisms responsible for colloid retention in porous media (from model simulations) and their fate (from experimental observations) in the presence of surface chemical heterogeneity under bulk unfavorable conditions.
This dissertation is divided into four projects that address current knowledge gaps concerning colloid transport:
The first project provides insight into the colloid-surface interaction energy when dissolved organic matter (DOM) is present in solution in the form of humic acid (HA). We investigated the origin and magnitude of opposing forces between silver and mica surfaces (representing nanosilver and sand grains) in solutions relevant to agricultural soils with direct measurements using a surface force apparatus. The results indicate that HA forms an adsorbed surface layer onto both surfaces with substrate-dependent properties. Ca$^{+2}$ significantly modifies the adsorption layer characteristics (thickness and compressibility), hence the interaction energy profiles. Force-distance measurements indicate that when silver-mica systems are exposed to HA, osmotic-steric, electrostatic and van der Waals forces dominate. Soft particle theory is deemed inappropriate for this system. We instead propose attachment efficiency estimates from measured surface properties, which suggest high particle mobility when nanosilver is applied to HA-rich agricultural soils with modest ionic strength.
The second project investigates the role that pore structure plays in colloid retention across scales with a novel methodology based on image analysis. High-resolution spatial profiles of retained particles from micro X-ray Computed Tomography allow quantification of the contribution from commonly proposed retention sites toward colloid immobilization. At the Darcy-scale, the spatial distribution of immobilized colloids along the porous medium reveals depth-independent partitioning of colloids among the pore-scale locations. The total mass of retained colloids exhibits non-monotonic deposition profiles, suggesting slow particle release from flow-stagnation zones. At the pore-scale, dominance and overall saturation of all retention sites indicates that the solid-water interface and wedge-shaped regions associated with flow-stagnation (grain-to-grain contacts in saturated and air-water-solid triple points in unsaturated conditions) are the greatest contributors toward retention under the tested conditions. At the interface-scale, xDLVO energy profiles are in agreement with pore-scale observations when the pore structure is taken into account. Our calculations suggest relatively favorable interactions for colloids and solid-water interfaces and for weak flocculation (e.g., at flow-stagnation zones), but highly unfavorable interactions between colloids and air-water interfaces. Overall, we demonstrate that pore-structure plays a critical role in colloid immobilization.
The third project demonstrates that aggregation of a electrostatically stable suspension is induced at typical groundwater velocities. Here, we compare the repulsive DLVO force between particle pairs to the hydrodynamic shear force opposing it acting on suspended particles. Column experiments imaged with high-resolution X-ray Computed Tomography are used to measure aggregate structure and describe their morphology, probability distribution, and spatial distribution. Distributions of aggregate volume and surface area are found to follow a power-law function. Aggregate Feret diameter is deemed to be exponentially distributed with some flow rate dependencies caused by erosion and restructuring by the fluid shear. Furthermore, size and shape of aggregates are heterogeneous in depth, where a small number of large aggregates control the shape of the deposition profile. The range of aggregate fractal dimensions implies a reaction limited aggregation process and a high potential for restructuring and/or breaking during transport. Therefore, while shear-induced aggregation is not currently considered in macroscopic models for particle filtration, it may be critical to consider in the processes that control deposition.
In the fourth project, the soil surface heterogeneity is found to extend the attachment efficiency and residence time of colloidal particles at the near surface. Both of these alterations significantly affect particle transport and retention at each relevant scale, such that macroscopic observations are deemed anomalous. At the interface-scale, we revise the interaction energy profiles considering chemical heterogeneities of the collector surface. Observations of deposited colloids at the pore scale are contrasted against continuum-scale predictions parameterized from rate coefficients upscaled from modeled Lagrangian trajectories. Results indicate that: i) Colloid-collector interaction energies are non-unique for surface heterogeneous systems and can span from strongly repulsive to strongly attractive; ii) The proportion of simulated particle by active retention mechanisms (fast/slow attachment and retardation at near surface) explain experimentally observed fraction of particles found at available pore-scale retention sites (solid-water interface and grain-to-grain contacts, respectively); and iii) Simulated non-exponential depth profiles of retained colloids and heavy-tailed breakthrough elution curves are in agreement with experimental mass balances and detailed depth profiles.
These collection of findings here reported have direct implications for the protection of subsurface ecosystems and water resources from potential contamination by hazardous colloids and for the rational management of agricultural soils. They are also critical to design strategies that effectively contain nanoparticle contaminant spreading in soils and groundwater.
