H23D-1615
Solute Mixing in Thermally Induced Buoyancy Flow
Lack of transverse mixing has been identified as a key factor limiting natural attenuation of continuously emitted contaminants, because transverse mixing controls reactions between the contaminants in the plumes and oxidants in the ambient groundwater. Since most plumes are rather flat, vertical mixing is particularly important. However, under normal groundwater flow conditions, vertical mixing is slow, even in heterogeneous media under transient flow conditions. We suggest enhancing vertical mixing by inducing buoyancy flow applying local injection of heat. This leads to fingers and rotational flow, increasing the contact surface between the plumes and the surrounding groundwater and thus increasing mixing. Several experiments were carried out in an insulated two-dimensional sandbox with a Pyrex glass front pane. The box was homogeneously filled with glass beads. All experiments were performed under water-saturated conditions. Two lines of a blue tracer dye were injected as the initial concentration. Electrical heaters were installed inside the sandbox as the heat sources. We tested two operation modes of the sandbox, one with horizontal and the other with vertical heating. Pictures in both the visible and infrared spectra were taken to measure the concentration and temperature distributions. Concentrations were computed by post-processing of the pictures taken in the visible spectrum, whereas the temperature distribution was obtained from the infrared pictures. The initial results showed that the heating created buoyant fingers and rotational flow, resulting in stronger dilution of the tracer. Simulations of the experiments using a modified version of SUTRA (USGS) agreed reasonably with the experimental data.
H23D-1616
Precipitation Of Iron Minerals In Porous Media
Formation of soil redoximorphic features, such as mottles, concretions and placic horizons is governed by Fe (II) and oxygen transport in porous media. We have conducted column experiments with four grades of silica sand that were placed vertically in the reduced saturated soil. Two sets of iron bands appeared within one week. Red oxidized bands (ferrihydrite and/or possibly lepidocrocite) were formed at the upper extent of the capillary fringe. Whereas, black bands (magnetite and/or green rust) were formed below the oxidized iron bands. These bands were transformed to the red bands as the experiment progressed. We propose a conceptual mechanistic framework for abiotic band formation and have developed a numerical model that incorporates the relevant factors contributing to the abiotic genesis of iron bands. This study reveals that abiotic iron oxide formation in unsaturated media is mainly a diffusion controlled process. The position of iron oxide bands is regulated by the diffusion rates of Fe (II) (aq) and Oxygen (g); thus elucidating the role of air-water and water-mineral interfacial properties, and diffusion gradients on iron transformations in natural systems. Our results suggest that the formation of Fe (III) oxide cemented bands in unsaturated hydrological systems is a rapid process occurring on a time scale relevant to human activity.
H23D-1617
Efficient Heat and Mass Transfer Formulations for Oil Shale Retorting
A mathematical model for oil shale retorting is described that considers kerogen pyrolysis, oil coking, residual carbon gasification, carbonate mineral decomposition, water-gas shift, and phase equilibria reaction. Reaction rate temperature-dependence is described by Arrhenius kinetics. Fractured rock is modeled as a bi-continuum consisting of fracture porosity in which advective and dispersive gas and heat transport occur, and rock matrix in which diffusive mass transport and thermal conduction occur. Heat transfer between fracture and matrix regions is modeled either by a partial differential equation for spherical conduction or by a linear first-order heat transfer formulation. Mass transfer is modeled in an analogous manner or assuming local equilibrium. First-order mass and heat transfer coefficients are computed by a theoretical model from fundamental rock matrix properties. The governing equations are solved using a 3-D finite element formulation. Simulations of laboratory retort experiments and hypothetical problems indicated thermal disequilibrium to be the dominant factor controlling retort reactions. Simulation accuracy was unaffected by choice of mass transfer formulation. However, computational effort to explicitly simulate diffusive mass transfer in the rock matrix increased computational effort by more than an order of magnitude compared with first-order mass transfer or equilibrium analyses. A first-order heat transfer approximation of thermal conduction can be used without significant loss of accuracy if the block size and/or heating rate are not too large, as quantified by a proposed dimensionless heating rate.
H23D-1618
Modeling cis-Dichloroethene Partitioning in Tetrachloroethene-NAPL Using Intra-aggregate Diffusion Sorption Models
Microbial facilitated dechlorination is being employed for treatment and polishing of source zones containing chlorinated ethenes. Metabolic reductive dechlorination converts tetrachloroethene (PCE) and trichloroethene (TCE) to lesser chlorinated ethenes and ultimately ethene. Where metabolic reductive dechlorination occurs in proximity to nonaqueous phase liquid (NAPL), the process may enhance the rate of NAPL dissolution by increasing the concentration driving force. Accumulation of cis-dichloroethene (cis-DCE) due to incomplete dechlorination, however, is commonly noted in reports of both laboratory- and field-scale studies of microbially enhanced dissolution. There is growing evidence that cis-DCE may partition into NAPL present within the source zone, thereby, enriching the organic phase with cis-DCE. Partitioning of cis-DCE into the NAPL may create a persistent source of cis-DCE contamination. A series of batch and column laboratory experiments undertaken to investigate this phenomenon provide a data set to investigate the fundamental processes controlling cis-DCE partitioning in PCE-NAPL. Analytical solutions previously developed to model sorption due to intra-aggregate diffusion were employed here to simulate column data. These models include aqueous-phase and non- aqueous phase resistance to mass transfer. Simulation results suggest that cis-DCE partitioning is dominated by the rate of diffusion within the entrapped NAPL ganglia, indicating that intra-NAPL diffusion may be a controlling mechanism for dechlorination rates observed in DNAPL source zones. Further insights were gained through a sensitivity analysis employing these analytical models. Preliminary conclusions from this study suggest that these rate limitations should be incorporated into multiphase compositional simulators used to describe metabolic reductive dechlorination.
H23D-1619
Scale-dependent kinetics of uranium desorption from contaminated subsurface sediments
Column experiments were performed to investigate the scale-dependent desorption of uranyl [U(VI)] from a contaminated sediment collected from the Hanford 300 Area at the US Department of Energy (DOE) Hanford Site, Washington. The sediment was a coarse-textured alluvial flood deposit containing significant mass percentage of river cobble. U(VI) was, however, only associated with its minor, fine-grained (< 2mm) mass fraction. U(VI) desorption was investigated both from the field-textured sediment using a large column (80 cm length by 15 cm inner diameter), and from its < 2mm, U(VI)-associated mass fraction using a small column (10 cm length by 3.4 cm inner diameter). Dynamic advection conditions with intermittent flow and stop-flow events of variable durations were employed to investigate U(VI) desorption kinetics and its scale dependence. A multi-component kinetic model that integrated a distributed rate expression with surface complexation reactions successfully described U(VI) release from the fine-grained, U(VI)-associated materials. The field-textured sediment in the large column displayed dual domain, tracer-dependent mass transfer properties that affected the breakthrough curves of bromide, pentafluorobenzoic acid (PFBA), and tritium. The tritium breakthrough curve showed stronger non- equilibrium behavior than did PFBA and bromide, and required a larger immobile porosity to describe. The dual domain mass transfer properties were then used to scale the kinetic model of U(VI) desorption developed for the fine-grained materials to describe U(VI) release and reactive transport in the field-textured sediment. Numerical simulations indicated that the kinetic model that was integrated with the dual domain properties best described the experimental results. The kinetic model without consideration of the dual domain properties over-predicted effluent U(VI) concentrations. Overall, our results indicated that the kinetics of U(VI) release from the field-textured sediment were different from that of its fine-grained, U(VI)-associated mass fraction. However, the desorption kinetics measured on the U(VI)-containing mass fraction could be scaled to describe U(VI) reactive transport in the contaminated field-textured sediment after proper consideration of the physical transport properties of the sediment. The research also demonstrated a modeling approach to integrate geochemical processes into field scale reactive transport models.
H23D-1620
MIXING-INDUCED PRECIPITATION: EXPERIMENTAL STUDY AND MULTI-SCALE NUMERICAL ANALYSIS
Laboratory experiments and pore scale and continuum scale simulations were used to study mixing-induced precipitation. In the laboratory experiment, sodium carbonate and calcium chloride solutions were injected at equal rates along two different halves of a quasi two-dimensional flow cell filled with quartz sand. Carbonate precipitation occurred within the mixing interface in a zone with uniform width of less than 5 mm. Pore-scale smoothed particle hydrodynamics simulations were conducted to study the mechanism of precipitation layer formation. SPH simulations were able to reproduce the precipitation event observed in the experiment. The simulations also revealed the presence of large pore-scale concentration gradients. This, and the presence of sub-continuum scale features such as a precipitation layer with characteristic width on the order of the average sand grain diameter, indicate the absence of a clear scale separation required for the existence of local continuum advection-dispersion equations. Nevertheless, we found that an adaptive high-resolution model based on traditional advection-dispersion equations with grid sizes in the mixing zone smaller than the size of the sand grains can also reproduce the essential features of the experiment such as formation of a thin precipitate layer of uniform thickness. As an alternative to high-resolution simulations, we proposed a new form of homogeneous and heterogeneous reactions terms in the advection dispersion equation. These terms involve transport and mixing indices that account for a non-uniform concentration distribution and highly localized reactions. The proposed model approach estimates the pore-scale distribution of concentrations due to homogenous and heterogeneous reactions during mineral precipitation where conventional low-resolution advection-dispersion equations produced erroneous results.
H23D-1621
Solute Generation Within Soil Pores Under Variably Saturated Conditions: A Simple Dimensionless Model
A major obstacle to understanding the mechanisms controlling weathering fluxes in the soil profile is the spatial variability of pore water chemistry. This chemical heterogeneity presumably reflects the diversity of soil micro- environments at the pore scale. Chemistry at the pore scale reflects each pore's structure, mineralogy, and water residence time. A mechanistic theory of solute chemistry at the scale of pedons, soil profiles, or entire catchments requires an understanding of how soils generate solutes at the pore scale, particularly under partially saturated conditions. We have developed a simple model to study how changes in matric potential alter rates of advection, diffusion and reaction of solutes within interconnected pores. A more realistic pore geometry than the standard capillary bundle model was used so that the role of thin films under laminar flow could be studied. In the model, mineral dissolution occurs along the walls of the pore, under both saturated and partially saturated conditions. The dimensionless Damkohler number (Da), which is the ratio of the reaction rate to the transport rate, and the Peclet number (Pe), which is the ratio of transport by convection to transport by diffusion, were calculated at each matric potential for a given pore size and geometry. These dimensionless numbers help us to identify the dominant mechanisms controlling solute concentrations under different soil moisture conditions.
H23D-1622
Effects of concentration fluctuations and gradients on reactive mixing
Sound understanding of mixing-controlled reactions in heterogeneous media is needed for the realistic modeling of contaminant transport in aquifers in the evaluation of natural attenuation processes, the design of nuclear waste disposal, and the engineered remediation of contaminated sites. Macroscopic formulations of reactive transport may not adequately describe mixing-controlled reactions because reaction rates are related to fine- grained concentration distributions. We use bimolecular dissolution-precipitation equilibrium reactions to demonstrate the segregation effects of reactants on the evaluation of effective reaction rates. The reaction rate predicted by macroscopic models is controlled by two competing effects: the large heterogeneity induced macrodispersion coefficient leads to an increase of reaction rate, while a more smoothed concentration gradient may lead to a decrease of the reaction rate. Under certain conditions, macroscopic models can give a good approximation at large times and away from the plume center of mass, due to the balanced variance budget, but should generally be viewed with skepticism. For example, they significantly underestimate reaction rate near the plume center due to the smoothed concentration gradient field.
H23D-1623
Characterization of Mixing and Spreading in a Bounded Stratified Random Medium
Matheron and de Marsily (1980) studied transport in a perfectly stratified infinite medium as an idealized aquifer model, for which they observed anomalous increase of the apparent longitudinal dispersion coefficient with the square root of time. Here, we investigate solute mixing and spreading in a vertically bounded stratified random medium. Large scale spreading and mixing is controlled by the dispersion time scale which measures the time for complete vertical mixing by local dispersion. Unlike for the infinite stratified medium, at asymptotically long times disorder-induced mixing and spreading is uniquely quantified by a constant "Taylor dispersion" coefficient. Preasymptotic transport, in contrast, is highly non-Fickian, characterized by the superdiffusive increase of the apparent dispersion coefficient. We focus on preasymptotic times and the transition to the asymptotic regime. The mechanisms that lead to enhanced solute spreading and mixing are studied in terms of effective dispersion coefficients that are derived from local moments, i.e., moments of the transport Green function (distribution density of a solute evolving from a point injection). We investigate spreading and mixing in single realizations as well as in an ensemble sense within a stochastic modeling framework. In the latter approach, the transport coefficients are defined as ensemble averages over all possible aquifer realizations.
H23D-1624
A Numerical Model of Coupled Fluid Flows, Thermal Transport, and Reactive Chemical Transport in Fractured and Porous Media
Accurate tools to reliably predict the migration and transformation of contaminants in the subsurface environment enhance the ability of environmental scientists, engineers and decision makers to analyze their impact and to evaluate the efficacy of alternative remediation techniques prior to incurring expense in the field. A mechanistic- based numerical model could provide such a tool. This paper communicates the development and verification of a mechanistically coupled fluid flow and reactive chemical transport under both fast and slow reactions in variably saturated porous and fractured media. Theoretical bases, numerical implementations, and modeling experiments of the model will be described. Two example problems will be presented. The first one will be a reactive transport problem to elucidate the non-isothermal effects on heterogeneous reactions. The second problem involves coupled fluid flows and reactive transport in fractured media. Its purposes are to examine the effects of precipitation and dissolution on fluid flow and matrix diffusion. It would help shade some lights on the retardation ability of grain matrices on the migration of metals and radionuclides.
H23D-1625
Investigation of Numerical Upscaling Techniques for Mixing-Controlled Reactions in Heterogeneous Media
Mixing of chemical species across plume boundaries has a major influence upon reactive pollutant fate in the subsurface. Small-scale heterogeneity leads to irregular plume boundaries which enhances mixing-controlled reactions through increasing the interfacial area of the plume. Therefore, it is crucial to capture this small-scale heterogeneity in order to properly model reactive transport. Unfortunately, computational limitations do not permit full resolution of the smallest scales of heterogeneity, and thus it is necessary to use a coarse numerical grid, particularly for cases with a large number of reactive species. In order to capture the sub-grid scale heterogeneity effects, we investigate the extension of multi-scale numerical techniques (which have been proved successful for diffusion and Darcy flow problems) to mixing-controlled reactive transport. The proposed upscaling technique is based on the multi-scale decomposition of the solution (which is similar to that proposed by Arbogast [1] for Darcy flow). We divide the governing system into two sub- problems – coarse-scale and fine-scale. We assume that the solute concentration has two components – coarse-scale (which is defined at the grid scale) and fine-scale (which is defined at the smallest modeled scale). The fine-scale sub-problems are solved locally by constructing numerical Green's functions, which are independent of the coarse-scale problem. The localization of fine-scale sub-problems is achieved by the closure assumption, which is enforced by prescribing appropriate boundary conditions for the fine-scale sub-problem. In this study, we investigate the effect of various closure approximations (i.e., boundary conditions for fine-scale sub- problem) on the overall accuracy of the numerical solution. We apply our multi-scale methods to several canonical problems for fast bi-molecular reactions. [1] T. Arbogast. Numerical subgrid upscaling of two-phase flow in porous media. In Z. Chen, R. E. Ewing, and Z.- C. Shi, editors, Numerical Treatment of Multiphase Flows in Porous Media, volume 552 of Lecture Notes in Physics, pages 35–49, Springer-Verlag, Berlin, 2000.
H23D-1626
Effects of flow focusing on enhancement of transverse-mixing limited reactions at the pore and continuum scale
For groundwater contaminants that originate from a persistent source, recent studies indicate that degradation occurs primarily along plume fringes, where limiting substrates and trace nutrients mix with the contaminant plume transverse to the direction of flow. In many cases, groundwater environments are heterogeneous, and plume fringes are located in sediments with large variations in permeability. Permeability variations can give rise to flow focusing, and a recent study [Werth et al., 2006] indicates that this can give rise to enhanced transverse mixing and reaction along plume fringes. While a number of studies have examined streamline focusing and enhanced mixing in individual pores, none have evaluated flow focusing at the pore scale, and subsequent enhancement of transverse-mixing limited reactions. Herein we present results from micromodel experiments and a pore-scale model that examine enhanced transverse mixing and chemical reaction in preferential flow structures. Micromodels are pore-scale representations of porous media etched into silicon wafers, and pore- scale modeling is performed using the lattice-Boltzmann method for flow, and a finite volume model for reactive transport. In both experiments and the model, two substrates are introduced into a network of pores via two separate and parallel fluid streams, and they mix due to transverse dispersion and react. In both the experiments and the model we control the exact porous-media structure, and therefore are able to directly evaluate how changes in preferential flow structure, mixing zone location, width of flow focused zone, length of focused zone, permeability contrast between high and low permeability zones, and intrinsic reaction rate affect mixing and product formation. Excellent agreement was obtained between the micromodel and LB-FVM results, indicating that the latter adequately captures the pore-scale behavior observed in the former. Results indicate that the extent of reaction for flow-focused scenarios increased by over 40% compared to identical structures not containing the flow focused region. Simulations evaluating the effect of flow focusing location on extent of product formation demonstrated that the longitudinal location does not have an effect on the overall extent of reaction for fast reactions, but does play a role when reaction rates are limiting. The length of the flow focusing region has a measurable but small effect on the overall extent of product formation. Comparison of pore-scale to continuum- scale results indicates that total product formation is similar, but that product formation at the pore-scale lags spatially behind product formation at the continuum-scale. Possible reasons for this will be discussed.
H23D-1627
UPSCALING REACTIVE TRANSPORT USING A PORE-NETWORK MODEL
The main objective of this research is a better understanding of the relations between the reaction rate constants and Darcy scale velocity using a 3D pore scale network model. First, we carry out numerical "experiments" in a single tube imposing a diffused concentration front moving into the tube and assuming equilibrium adsorption at its wall. We have found that the form of volume-averaged concentration breakthrough curves can be described only by a linear kinetic model. The numerical data are used to determine the rate constants of the linear kinetic adsorption equation as functions of tube geometry, subscale adsorption parameters, and average flow velocity. Next, we have constructed a 3D pore-network model which is composed of a large number of interconnected tubes. Transport equations for kineticly-adsorbed solutes are solved within each tube. Single tube results are used in these equations. The pore-network model is used to simulate core-scale solute transport. Numerical data are used to calculate flux-averaged concentration to determine upscaled kinetic rate coefficients. In this way, we have been able to develop a relationship between core-scale adsorption rate constants and local-scale equilibrium coefficient and average flow velocity.
H23D-1628
Multiscale Reductions in Mass Flux as a Function of Residual Saturation
Remediation strategies for reducing nonaqueous phase liquid (NAPL) mass have in many cases has been successful in removing significant portions of the contaminant mass originally present in the system. However, significant reductions in entrapped NAPL mass may not lead to sufficiently low mass fluxes from the source zone to meet remediation goals in the short-time limit. We consider source zone mass flux as a function of NAPL residual saturation for systems ranging from small scale laboratory studies to a pilot scale field investigation. We compare the residual saturation source zone mass flux relationship across this wide range of scales and compare to other observations made in the literature. We make predictions of the laboratory experiments based upon pore-scale modeling using only the grain size distribution, porosity, and NAPL characteristics as input. We simulate the velocity field using Lattice Boltzmann methods and model the dissolution process using high resolution numerical methods at the pore scale. We compare the simulation results with multiscale experimental observations and show the importance of NAPL morphology. We also comment on both the benefits and potential shortcomings of active NAPL source-zone remediation for meeting typical environmental cleanup standards and summarize realistic expectations for active source-zone remediation.
H23D-1629
Comparison of Models Used to Evaluate Mass Removal and Mass Flux Reduction
The purpose of this study was to investigate the application of models of varying complexity to the dissolution of non-uniformly distributed immiscible liquid in physically heterogeneous systems at both the intermediate and field scale. Flow cell experiments focused specifically on characterizing the relationship between mass flux reduction and mass removal for systems wherein immiscible liquid is poorly accessible to flowing water. Both end-point and time continuous data from several field studies were examined to evaluate observed relationships between mass flux reduction and source-zone mass removal. Methods for estimating mass-flux- reduction/mass-removal behavior, based on the use of simple mass-removal functions and 1-D and 3-D mathematical flow and transport models, were applied to the measured data. The simple mass-removal function generated singular curves that could not reproduce the multi-step behavior exhibited by data from both laboratory and field studies. The permeability field and the distribution of the immiscible-liquid zones were represented explicitly in the 3-D model. In contrast, the system was conceptualized as a pseudo-homogeneous medium, with immiscible liquid uniformly distributed throughout the system for the 1-D modeling. With this approach, all factors that influence immiscible-liquid dissolution are incorporated into the calibrated dissolution rate coefficient, which in such cases serves as a composite or lumped term. The calibrated dissolution rate coefficients obtained for the 1-D modeling were approximately two to three orders of magnitude smaller than the values obtained from column experiments, and which were used for the 3-D modeling. The disparity in magnitudes of the values used for the 1-D and 3-D modeling reflects the difference in implicit versus explicit consideration of the larger-scale factors influencing immiscible-liquid dissolution in the systems. However, the calibrated dissolution rate coefficients were similar for the various flow-cell experiments, which suggest that the model was relatively insensitive to the specific nature of the source-zone configuration.
H23D-1630
Experimental Study of U(VI) Release Kinetics from Aquifer Sediments from a Former Uranium Mill Tailings Site (Rifle, Colorado, USA)
Uranium(VI) release kinetics from aquifer sediments from a former uranium mill tailings site in Rifle, Colorado was studied to understand uranium distribution within the sediments. The sediments were sampled at depths of 3.5-3.8 m in December 2004. The samples were air-dried, sieved, and the <2 mm fraction was collected and used in this study. Total uranium content in the sediments, determined by gamma-radiometry, was 4.1 μg/g sediment. The labile fraction of U(VI) in the sediments was determined using carbonate/bicarbonate extractions, which should cause complete desorption of U(VI) in the absence of mass transfer limitations. Carbonate/bicarbonate extraction of the sediments showed very slow release kinetics, with only 12 % of the labile U(VI) in the sediments being released during the first 96 hours of extraction. This is much less than found in a previous study at a different mill tailings site (Naturita, Colorado), in which more than 80 % of labile U(VI) was released during the same period of extraction. Up to two months of carbonate/bicarbonate extraction released 1 μg U(VI) per gram of Rifle sediment, which is 25 % of the total U in the sediment. Extraction with an artificial groundwater prepared to simulate the field groundwater chemistry showed 0.26 μg U/g sediment was released during the initial 94 hours of extraction, with a gradual increase of released U(VI) with time, while other major and minor elements (except Si) rapidly reached steady-state concentrations during the first few hours of reaction. Two hypotheses are under consideration to explain the slow U(VI) release kinetics: 1) colloidal clay fraction particles cementing larger grains of the sediments are creating nanoscale interparticle pores that act as a diffusion barrier to U(VI) desorption, and 2) a U(IV) solid phase exists whose oxidation and dissolution control the U(VI) release rate. To test the hypotheses, oxidation and extraction of the sediments have been conducted using oxidants such as hydrogen peroxide. The results of this study are expected to contribute to better understanding of U(VI) sorption behavior on the sediments and will contribute to the development of a comprehensive bioremediation strategy at the Rifle uranium mill tailings site.