H51G-0851
Transport and Fate of Titanium-Dioxide Nanoparticles in Heterogeneous, Unsaturated Porous Media
The increased production and use of manufactured nanoparticles presents concerns for human and environmental health, especially in light of the recently discovered cytotoxicities of many nanoparticles. Nanoparticles can exhibit unique chemical and physical properties compared to their bulk counterparts, which may influence their mobility in the subsurface. Previous studies have shown that the vadose zone plays an important role in solute transport, and may entrap colloids at the air/water interface. Aquifer recharge may release colloids from the vadose zone into the saturated zone, where the potential for mobility is greater. To date, little information has been reported regarding the exchange of nanomaterials between the vadose and saturated zones. Recent laboratory experiments have examined the effects of macro-heterogeneity in the vadose zone on the transport of titanium-dioxide nanoparticles. A custom designed, large-scale, one-dimensional vertical column has been used to study natural infiltration and drainage of nanoparticle suspensions in unsaturated soils. It has been observed that the spatial configuration of the hydraulic properties of a heterogeneous soil profile can dramatically influence the mass of titanium-dioxide nanoparticles retained in the vadose zone. Results of experiments and corresponding flow and transport model simulations will be discussed, with an emphasis on the implications for nanomaterial mobility and associated risk.
H51G-0852
Impact of Colloid Size on its Transport in Porous and Fractured Media
The physicochemical factors affecting the transport of colloids in the subsurface are of foremost interest from both scientific and applicable points of view. One of the most important parameters is the colloid size. The transport of three sizes of latex microspheres, 0.02, 0.2 and 1 micrometer was explored in: (1) homogenous silica sand following three different levels of treatment; (2) unsaturated homogeneous sand; (3) unsaturated heterogeneous sand, with three levels of water content; and (4) two natural discrete fractures crossing chalk cores, with equivalent hydraulic apertures of 180 and 350 micrometer. In all case it was found that maximum recovery was obtained for the 0.2 micrometer microspheres, slightly lower recovery for the 1 micrometer colloids, and much lower recovery for the 0.02 micrometer colloids. Retention of the 0.02 micrometerƒncolloids was more sensitive to the level of sand cleaning than that of the larger colloids. In the natural chalk fractures, recovery of the 0.2 and 1 um colloids ranged between 79-99 percent while for the 0.02 micrometer colloids recovery was around 75 percent. The small colloids retained within the fractures could not be remobilized even under relatively fast flow rates of the dispersive solution, therefore it was concluded that they irreversibly penetrate the complex porous media of the surrounding chalk matrix (average pore size 0.15 micrometer). A comparison between transport through fractures of dense clay particles and buoyant microspheres, within the same size range, reveals that in addition to size, colloid density also plays an important role. Interestingly, despite the different mechanisms that play the dominant role for colloid retention depending on the media collector properties, the common observation is maximum recovery of middle-size colloids (0.2 micrometer) less recovery for larger colloids, and minimum recovery for the smallest colloids.
H51G-0853
Effect of Nanoparticle Aggregation, Polydispersity, and Concentration on Transport of Surface- Modified Nanoscale Zerovalent Iron (NZVI) Particles in Saturated Porous Media
Nanoscale zerovalent iron (NZVI) particles for in situ subsurface remediation are typically injected at high particle concentration (1-10 g/L) to minimize costs. At this high particle concentration, aggregation and media ripening effects might limit transportability of the particles. Fundamental understanding of the phenomena controlling the deposition and transport of concentrated surface modified NZVI dispersion is needed for efficient delivery and placement in the contaminant source zone. This study investigates the role of aggregation, particle polydispersity, and particle concentration in transport of poly(styrene sulfonate) (PSS) modified NZVI in saturated sand columns. Bench-scale column experiments are performed at various particle concentrations (0.03 to 6 g/L) at low but environmentally relevant ionic strength, 10 mM Na+, a pore water velocity of 3.2 x 10-4 m/s, and an average collector size of 300 ìm. To elucidate the importance of particle polydispersity, transport and deposition of PSS-modified NZVI with three different particle size distributions are compared. The influence of intrinsically magnetic particle-particle interaction of PSS-modified NZVI on their aggregation and deposition in porous media is assessed by comparing the deposition behavior of PSS-modified NZVI (magnetic) with PSS-modified hematite (nonmagnetic) of the similar surface properties. The transport of PSS70K-modified hematite is not sensitive to particle concentration (from 30 mg/L to 6g/L) or particle polydispersity. In contrast, at high particle concentration (1 to 6 g/L), the transport of PSS70K-modified NZVI (magnetic particles) is sensitive to particle polydispersity, but insensitive to particle concentration. Because the adsorbed layer properties of PSS70K-modified NZVI provide roughly the same particle-collector interaction energies, the difference in deposition behavior of the different NZVI size fractions is attributed to particle-particle interaction (aggregation) which is significantly different between three different size fractions of PSS70K-modified NZVI because magnetic attractive forces increase with r6. This study emphasizes an important role of particle-particle interaction (aggregation) and particle polydispersity in the transport of concentrated NZVI dispersion in porous media.
H51G-0854
Mobility of Multi-walled Carbon Nanotubes in Porous Media
Engineered multi-walled carbon nanotubes (MWCNTs) are the subject of intense research and are expected to gain widespread usage in a broad variety of commercial products. However concerns have been raised regarding their potential environmental and health risks. The mobility of MWCNTs in porous media is examined in this study through one dimensional flow-through column experiments under conditions representative of subsurface and drinking water treatment systems. The goal of this work was to determine dominant MWCNT removal mechanisms and factors that control MWCNT transport. Results demonstrate that pore water velocity strongly influenced MWCNT transport, a result that stands in contrast to traditional colloid filtration theory, which suggests a relatively minor effect of flow velocity in comparison to Brownian diffusion. Experiments conducted at different ionic strengths indicate that both particle deposition and straining are important MWCNT removal mechanisms from the aqueous phase. Given these findings, traditional colloid filtration theory may not be appropriate for the prediction of MWCNT mobility in porous media. This may be due to the large aspect ratio of the MWCNTs and the importance of straining in MWCNT removal.
H51G-0855
Refining Colloid Filtration Theory for Nanoparticle Transport in Porous Media
Classical colloid filtration theory (CFT) was developed in the 1970's for predicting the rate of particle removal in deep-bed filtration water treatment systems. Over the past two decades CFT has been widely applied to address subsurface colloid transport research questions, but current concerns on the environmental risks of nanoparticles coupled with the uncertainties of CFT validity for nanoscale colloids calls for development of new constitutive theory for nanoparticle transport. The premise of CFT is that macroscopic transport can be inferred from a mechanistic analysis of particle transport within a single pore space. It has been demonstrated recently by two independent studies (Tufenkji and Elimelech, TE; Nelson and Ginn, NG) that the classic Rajagopalan and Tien (RT) formula significantly overestimates the collector efficiency (rate of contact with porous media surface) for submicron particles due to its assumption of independence between Brownian motion and the other components of particle motion. However, the NG study suggests that the RT equation overestimates the collector efficiency for all submicron particles while the new TE equation converges with RT as particle size decreases (i.e., enters the nanoparticle range). We hypothesize that the reason for this discrepancy is an uncaptured particle size effect on hydrodynamic retardation and van der Waals forces. We present preliminary results aimed at testing this hypothesis and providing a new collector efficiency equation that is more robust for predicting the transport of nanoparticles in porous media.
H51G-0856
The Stability of Cerium Oxide Nanoparticles.
Interest has been shown in cerium oxide nanoparticles due to their potential use catalysts, gas sensors, and more. Unchecked use of these nano-materials could have profound environmental impacts, however, little is known about the fate and transport of these nanoparticles in aquatic systems. In this work, small (d ~ 15 nm) mostly spherical cerium oxide nanoparticles were synthesized in house and observed using transmission electron microscopy. The zeta potential of the particles was also measured across a range of pH values to determine the point of zero charge. Dynamic light scattering measurements demonstrate that these particles exist as supra-aggregates in solution and not as individual nanoparticles. The aggregation kinetics of cerium oxide supra-aggregates were measured across a range of sodium chloride concentrations using time resolved dynamic light scattering. Stability diagrams were constructed by comparing the measured aggregation rate constants to diffusion controlled rate constants. The stability of cerium oxide aggregates was modeled using conventional DLVO theory to determine a Hamaker constant for cerium oxide. The system showed significant deviations from DLVO theory at low ionic strengths where the supra-aggregates were less stable than predicted. The origins of these deviations will be discussed.
H51G-0857
Transport of engineered zeolite and natural nanoparticles in porous media
There are many natural nanoparticles (NPs) that are ubiquitous in the environment such as soil and sediment colloids. In addition, many new engineered NPs, such as tailored zeolites, are being developed for applications in which they may be released into the environment. The fate and transport of the NPs is very much related with contaminant fate and transport. This study focused on transport of engineered zeolite nanoparticles (NPs) and natural soil and sediment colloidal NPs within porous media under saturated conditions. Clean medium-sized sand grains were used as the porous media and NPs were injected into the column as a pulse. KCl or CaCl2 with varying concentrations was used as background electrolyte. The results showed that, interestingly, the zeta- potential of the natural colloids and Zeolite-Ca decreased (more negative) with increasing KCl concentration while increased (less negative) with increasing CaCl2 concentration. This unexpected results was attributed to the fact that the natural colloids and Zeolite-Ca are saturated with divalent cations (Ca2+ and/or Mg2+) originally and the replacement of these divalent cations with K+ on the colloid surfaces caused the zeta-potential to drop with increasing KCl concentrations. The zeta-potential measurement of Zeolite-K increased with either KCl or CaCl2 concentration. Consistently early breakthrough was observed for NP compared with conservative tracers (KCL or CaCl2) and the effect was more pronounced with higher water flowrate. Zeolite-K showed significantly higher degree of transport (defined as percent of NPs transported out of the column) than Zeolite-Ca under the otherwise same conditions. With KCl as the background electrolyte, the significantly higher NP transport was observed than with CaCl2. Overall, as the ionic strength of the flowing fluid increased, the transport of the NPs decreased, largely due to the compressed double layer under the higher ionic strength. Besides, as the flow rate of the flowing fluid increased, the transport of the NPs increased due to less interaction time between the NPs and the porous media. Analysis of NP transport data showed that Zeolite-Ca fell into the same category as natural colloids while Zeolite-K deviated significantly from the others, suggesting the cations on the cation exchange capacity (CEC) sites play a large role in NP transport. Our results also revealed that the CEC/(surface area) can be a good indicator for predicting relative transport of the NP within the same group. An empirical model was developed to predict NP transport under the experimental conditions. The model parameters include flow rate, ionic strength of the solution, charge density (CEC/surface area). In sum, for the hydrophilic colloids, the chemistry (such as cation species) on the colloids surfaces and flowing fluid is important for predicting their transport in the environment.
H51G-0858
Stress/Strain Cycling Effects on Colloid Transport in a Synthetic Porous Medium
Laboratory experiments on porous rock cores have shown that seismic-band (100 Hz or less) mechanical stress/strain cycling of the rock matrix can mobilize sub-pore-size particles (colloids) trapped in the pore space and allow them to be expelled during steady-state water flow. This coupling of dynamic stress to colloid mobility is a potential key mechanism whereby seismic waves may alter formation permeability and porous mass transport in Earth's crust. Prior experiments where colloid suspensions were injected into Fontainebleau sandstone demonstrated that colloid size and the ionic strength of the suspending fluid are major parameters that will control the ability of the colloids to attach to pore walls or to form particle bridges at pore throats. Both effects can lead to significant changes in permeability. A unique core-holder apparatus that applies low-frequency mechanical stress/strain to 2.54-cm-diameter porous rock samples during constant-rate fluid flow was used for those experiments. It was found that dynamic stress at 25 to 50 Hz could mobilize these trapped colloids only when the ionic strength is low. These earlier experiments on natural rocks were difficult to interpret in terms of how the colloids distributed themselves throughout the heterogeneous pore space and what interactions were occurring between the colloids and the solid matrix. Observed permeability changes appeared to be confined to the first 5-10 cm of the rock where the colloids were injected, yet significant transport of colloids was observed along the entire length of the sample. The "natural rock" system is too complex geometrically at the pore scale to allow quantification of mass transport properties along its entire length. To remedy this problem, new colloid transport experiments were performed with a synthetic glass-bead pack. Unconsolidated 1-mm-diameter borosilicate beads were packed into a confining sleeve to a length of 30 cm. Sufficient radial and axial confinement pressures were applied to the sleeve to create a fluid seal and to introduce rigidity to the sample. Suspensions of 2-um polystyrene microspheres in deionized water were injected into the bead pack while maintaining constant flow through the sample. Sample permeability and effluent microsphere production were measured before, during, and after low-frequency stress cycling, as in the previous experiments on rocks. Because the core is composed of uniform sized beads, the detailed microsphere breakthrough curves combined with precise permeability measurements as a function of position along the bead pack allow identification of regions where applied dynamic stress has the largest effects on colloid mobility. Preliminary results of these new experiments will be presented.